why banks are not allowed in bankruptcy - school of law

985

Why Banks Are Not Allowed in Bankruptcy

Richard M. Hynes* Steven D. Walt**

Abstract

Unlike most other countries, the United States uses different procedures to resolve insolvent banks and nonbank firms. The Bankruptcy Code divides control over nonbank firms among the various claimants, and a judge supervises the resolution process. By contrast, the FDIC acts as the receiver for an insolvent bank and has almost complete control. Other claimants can sue the FDIC, but they cannot obtain injunctive relief, and their damages are limited to the amount that they would have received in liquidation. The FDIC has acted as the receiver of insolvent banks since the Great Depression, and the concentration of power in the FDIC is traditionally justified by two arguments: (i) FDIC control speeds the disposition of the bank’s assets which maintains the liquidity of deposits and encourages faith in the banking system, and (ii) the FDIC’s role as the largest creditor gives it an incentive to maximize recovery. We ask whether they still (or ever did) justify FDIC control. The first argument fails because it conflates the need for a timely satisfaction of the claims of insured depositors by the FDIC with the need to quickly dispose of the failed bank’s assets. As stated, the second argument fails to justify FDIC control as the largest creditor can take self-interested actions harmful to other claimants. However, the FDIC is not merely the largest creditor. A detailed survey of the capital structure of failed banks reveals that the FDIC is usually the only major creditor and that the value of the FDIC’s claim nearly always exceeds the value of a failed bank’s assets. The FDIC is the residual claimant and has the incentive to make the right decisions in disposing of the bank’s assets. We question whether this principle can justify recent legislation that extends FDIC control over the resolution of large bank holding companies. We further consider

* Professor of Law, University of Virginia School of Law. ** Percy Brown Jr. Professor of Law, University of Virginia School of Law. We thank George Geis, Ethan Yale and participants at workshops at Northwestern University and the University of Virginia for valuable comments. We thank Jon Ashley, Dane Dunson, Benjamin Grosz, and Kent Olsen for valuable research assistance. All errors remain our own.

986 67 WASH. & LEE L. REV. 985 (2010)

four limits on our argument for FDIC control: (i) capital structure is endogenous—the absence of claims junior to the FDIC may reflect the lack of voice given to these claimants in a bank resolution process, (ii) agency costs internal to the FDIC may prevent the FDIC from maximizing the recovery from the failed bank’s assets, (iii) the FDIC may not be the residual claimant of extremely large banks with complex liability structures, and (iv) debt conversion schemes which allow for automatic financial restructuring of a failed bank may render bank resolution procedures less necessary. This Article argues that these limits do not justify removing the FDIC from control in resolving most bank failures.

Table of Contents

I. Introduction ...................................................................................... 986

II. Bankruptcy and Bank Insolvency.................................................... 993 A. Dispersed Control in Bankruptcy ............................................. 993 B. Bank Insolvencies and the Purchase and Assumption Agreement ............................................................ 998

III. The Case for FDIC Control of Bank Insolvencies ........................ 1006 A. The Benefits of Speed and Secrecy ........................................ 1007 B. The FDIC as Residual Claimant ............................................ 1012 1. The FDIC as the Largest Creditor ................................... 1016 2. FDIC Losses on Failure ................................................... 1021

IV. Limits of the Case for FDIC Control............................................. 1025 A. Agency Costs in Bank Resolutions ........................................ 1027 B. The Liability Structure of Large Banks ................................. 1031 C. Self-Financed Restructuring: Debt Conversion .................... 1036

V. Bank Holding Companies and Nonbank Financial Companies ...................................................................................... 1044

VI. Conclusion ..................................................................................... 1051

I. Introduction

Most nations resolve failed banks with the same procedures they apply to other insolvent firms.1 American law is different. American banks and 1. See Rosalind L. Bennett, Failure Resolution and Asset Liquidation: Results of an

WHY BANKS ARE NOT ALLOWED IN BANKRUPTCY 987

thrifts do not receive bankruptcy protection.2 Instead, regulators seize insolvent or unsound banks or thrifts and give the Federal Deposit Insurance Corporation (FDIC) the authority to resolve them. Almost always the FDIC chooses to resolve seized institutions through a receivership.3 Very different rules govern the bankruptcy and bank receivership processes. These rules appear in different titles of the United States Code and have important substantive differences. The most important difference between the two procedures is the concentration of control over the disposition of the failed firm’s assets. The traditional bankruptcy reorganization divides control among the various claimants and appoints a judge to supervise the process. The overwhelming majority of reorganizations are resolved consensually with the approval of each class of creditors and shareholders.4 Even when a debtor tries to "cram down" a plan over the objections of dissenting creditors, the debtor must win approval of at least some creditors, and the other creditors can ask the judge to reject the plan because it fails to comply with tests of horizontal and vertical equity or it is not in the best interests of the creditors.5

Debtors sometimes bypass the traditional reorganization with a relatively quick sale of many of the failed firm’s assets or at least the assets necessary for some of the business to continue operating. Although this

International Survey of Deposit Insurers, 14 FDIC BANKING REV. 1, 9 (2001) ("Outside the United States, most failed banks go through a regular corporate bankruptcy process."). 2. See 11 U.S.C. §§ 109(b), (d) (2006) (stating that banks are ineligible for bankruptcy, so that neither the bank nor the bank’s creditors can place the bank in bankruptcy). As noted below, bank holding companies can file for bankruptcy in the United States, and many of the largest bankruptcies on record have been bank holding companies. See infra note 31 and accompanying text (discussing the Washington Mutual, Inc. bankruptcy). 3. The FDIC can decide to resolve the bank by a conservatorship or by a receivership. Conservatorships are very rare, however. Between 1934 and 2005 only two banks were resolved in conservatorships. See RICHARD S. CARNELL, JONATHAN R. MACEY & GEOFFREY P. MILLER, THE LAW OF BANKING AND FINANCIAL INSTITUTIONS 706 (4th ed. 2009) (discussing conservatorships as a mechanism to resolve a bank). For convenience, we therefore use the term "bank receivership" to refer to any proceeding designed to resolve an insolvent bank. 4. See Arturo Bris, Ivo Welch & Ning Zhu, The Costs of Bankruptcy: Chapter 7 Liquidation Versus Chapter 11 Reorganization, 41 J. FIN. 1253, 1256, 1274 (2006) ("Although the court can use ‘cram down’ to pass the plan and save time, no court used this in our sample."); Lynne M. LoPucki & William C. Whitford, Bargaining Over Equity’s Share in the Bankruptcy Reorganization of Large, Publically Held Companies, 139 U. PA. L. REV. 125, 141 (1990) (recognizing an extraordinarily high level of settlement). 5. See 11 U.S.C. § 1129(b) (outlining the requirements for judicial approval of a plan).

988 67 WASH. & LEE L. REV. 985 (2010)

quick sale denies the creditors the ability to vote on the sale, the debtor must still seek approval of the bankruptcy judge,6 and the dissenting creditors can at least ask the judge to enjoin the sale. The recent Chrysler bankruptcy provides a nice illustration of this process. Within a few weeks of Chrysler’s bankruptcy filing, the debtor sold substantially all of its operating assets to a "New" Chrysler owned by the UAW, Fiat and the U.S. and Canadian governments.7 New Chrysler emerged from bankruptcy and continued manufacturing and selling cars.8 The original corporate entities, and the few assets that New Chrysler did not want, stayed in bankruptcy, and Old Chrysler remains there over a year after its bankruptcy filing.9 Eventually the bankruptcy process will conclude, and these remaining assets and the proceeds of the sale to New Chrysler will be distributed to the creditors of Old Chrysler. This sale occurred outside of a plan of reorganization, and the creditors did not vote on the sale. However, a bankruptcy judge did approve the sale, and the creditors had a right to, and did, argue that the approval should have been denied. Apart from its notoriety, the way in which Chrysler’s principal assets were disposed of now is fairly typical in large bankruptcy reorganizations.10

A bank receivership begins when the FDIC seizes control of the bank. In the vast majority of cases, the FDIC identifies an acquiring financial institution before it seizes the failed bank.11 The acquirer will assume some or all of the failed bank’s deposits and perhaps some of the failed bank’s other liabilities as well. As consideration, the acquirer will receive some of 6. See id. § 363 (regulating the use, sale, or lease of property by a trustee). 7. See Mark H. Anderson & Jeff Bennett, Pension Funds Ask High Court to Delay Sale of Chrysler, WALL ST. J., June 8, 2009, at B3 (discussing the Chrysler sale and the action taken by junior creditors to stop the sale). 8. Id. 9. See Chrysler Restructuring Homepage, http://www.chryslerrestructuring.com (last visited Sept. 25, 2010) (providing information about the restructuring of Chrysler) (on file with the Washington and Lee Law Review). 10. See CONGRESSIONAL OVERSIGHT PANEL, SEPTEMBER OVERSIGHT REPORT: THE USE OF TARP FUNDS IN THE SUPPORT AND REORGANIZATION OF THE DOMESTIC AUTOMOBILE INDUSTRY (2009) (providing an overview of the Chrysler reorganization). Two features of Chrysler’s sale were somewhat unusual: The sold assets remained subject to certain liabilities and interests, and the bidding was restricted to offers that assumed these liabilities and interests. See id. ("[A]ssets can under appropriate circumstances be sold ‘free and clear’ in a 363 sale, but that in Chrysler the buyer . . . took the assets subject to specified obligations to the UAW Trust."). However, the timing and judicial approval of the sale outside a reorganization plan were typical. 11. See infra Figure 1 (providing statistics on the method of resolution chosen by the FDIC). The FDIC had identified an acquirer in every purchase and assumption ("PA" or "PI") transaction.


the failed bank’s assets and will nearly always receive a payment from the FDIC. The FDIC usually retains most of the failed bank’s assets12 and takes months or even years to liquidate these assets.13 However, the initial purchase and assumption occurs much more quickly than the initial sale of a firm’s assets in bankruptcy. The automobile bankruptcies generated headlines because the sales were completed in a matter of weeks. A typical purchase and assumption of a failed bank is quicker. The FDIC nearly always seizes a failed bank at the close of business on a Friday,14 and some of the failed bank’s offices may reopen as part of the acquiring bank the following Monday.15

The FDIC enjoys a level of control that a dominant creditor could only dream of obtaining in bankruptcy. For one thing, the initial purchase and assumption will be complete long before a judge can conceivably hear an objection. Even if a creditor’s objection is timely, the law significantly restricts the grounds upon which a creditor can complain about the sale. For instance, courts generally find that applicable law does not allow the creditor to complain that an alternative sale would have fetched a higher price.16 The 12. See infra note 81 and accompanying text (discussing the amount of assets of the failed bank that the acquirer acquires). 13. See DIV. RESOLUTION & RECEIVERSHIPS, ASSET DISPOSITION MANUAL 22 (1999) (providing a timeline for the disposition of a failed bank’s assets). 14. See FDIC: RESOLUTIONS HANDBOOK 16 (2003) ("The chartering authority closes the institution and appoints the FDIC as receiver (usually on a Friday)."). The seizure of Washington Mutual was a rare exception to this rule. See E. Scott Reckard & Tiffany Hsu, Feds Seize, Sell WaMu in Biggest U.S. Bank Failure, L.A. TIMES, Sept. 26, 2008, at A1 ("In the biggest bank failure in U.S. history, Washington Mutual Bank was seized late Thursday by federal regulators and immediately sold . . . ."). 15. See Reckard & Hsu, supra note 14, at A1 (noting that the transition would be "seamless" and that there would not be any interruption in services provided to customers). A typical example is the recent failure of IndyMac Bank. See Kathy M. Kristoff & Andrea Chang, Federal Regulators Seize Crippled IndyMac Bank, L.A. TIMES, July 12, 2008, at A1 ("The federal government took control of Pasadena-based IndyMac bank on Friday in what regulators called the second-largest bank failure in U.S. history."). The bank’s thirty-three branches were scheduled to reopen as branches of the acquirer (IndyMac Federal Bank) with the resumption of normal business hours. See id. ("Normal branch hours, online banking and phone banking services are to resume Monday."). Although the purchase and assumption is resolved within days of the initiation of the receivership, the FDIC typically spends about three months soliciting bids and preparing for the receivership. See John R. Walter, Closing Troubled Banks, 90 ECON. Q. 51, 58 (2004) ("The FDIC then begins a multistep process generally lasting 90 to 100 days, but which can proceed much more quickly."). Similarly, insolvent nonbank firms can spend a great deal of time preparing for their bankruptcy filing in order to minimize the duration of the proceedings. 16. See Mosseri v. FDIC, 2001 U.S. Dist. Lexis 18899, at *14 (S.D.N.Y. 2001) ("[C]ongress did not intend to create a private right of action for disappointed bidders for FDIC assets under 12 U.S.C. § 1821(d)(13)(E)(i)–(iii)."); 19 Court St. Assocs., LLC v.

990 67 WASH. & LEE L. REV. 985 (2010)

law also limits complaining creditors to a monetary remedy based on the amount that they would have received in a liquidation of the bank’s assets. Further, a court cannot enjoin the FDIC’s sale of the failed bank’s assets.17

Articles in the legal literature discuss the initiation of the bank resolution process.18 But there is surprisingly little discussion of what happens after the process has been initiated. This Article focuses on the resolution process itself. The literature offers two possible reasons for giving the FDIC the extraordinary control it has. The first is speed.19 Control gives the FDIC the ability to sell assets with no interference from other creditors and little judicial oversight. This allows for quick resolutions. A swift resolution, in turn, can ensure liquidity for depositors, increase depositors’ confidence in the banking system, and prevent runs on both failed and healthy banks. We are not persuaded by this justification. The identified benefits of speed depend on the timely reimbursement of insured depositors (a matter governed by the terms of FDIC insurance) and not on the sale of the failed bank’s assets. In fact, the FDIC’s own liquidation manual does not call for the completion of the asset disposition until four years after the seizure of the bank.20 The FDIC itself recognizes that disposing of all of the failed bank’s assets takes time.

The second reason for FDIC control is that it is likely the largest creditor of the failed bank and therefore has an incentive to maximize the recovery from the bank’s assets.21 But bankruptcy does not grant the largest creditor

Resolution Trust Corp. (In re 19 Court St. Assocs., LLC), 190 B.R. 983, 994 (Bankr. S.D.N.Y. 1996) ("[A]s a disappointed bidder, Court Street does not have a private right of action against the RTC for the alleged violation of the sections dealing with fair and consistent treatment of offerors and with maximization of net value return on assets."). 17. See 12 U.S.C. § 1821(j) (2006) ("Except as provided in this section, no court may take any action, except at the request of the Board of Directors by regulation or order, to restrain or affect the exercise of powers or functions of the Corporation as a conservator or a receiver."). 18. See, e.g., David F. Skeel, The Law and Finance of Bank and Insurance Insolvency Regulation, 76 TEX. L. REV. 723, 724 (1998) (discussing initiation issues and proposing possible reforms); Peter P. Swire, Bank Insolvency Law Now That It Matters Again, 42 DUKE L.J. 469, 477–90 (1992) (discussing the nature of special bank insolvency law); see also Douglas G. Baird, The Initiation Problem in Bankruptcy, 11 INT’L REV. L. & ECON. 223, 223–33 (1991) (discussing the optimality of the procedure for initiating a bankruptcy). 19. See, e.g., Bennett, supra note 1, at 7 ("The liquidation system is governed by receivership laws that seek to ensure the speedy resolution of banks and that therefore allow the receiver broader powers than the bankruptcy laws allow."). 20. See supra note 13 and accompanying text (providing a timeline for the disposition of a failed bank’s assets). 21. See, e.g., Bennett, supra note 1, at 9 ("When Congress created the FDIC, it believed that making the largest creditor (the FDIC) responsible for liquidating the assets of


control over the resolution of nonbank firms. All claimants (including shareholders) are given a voice to ensure that the largest creditor does not use its control to their detriment. However, the available evidence suggests that the FDIC is far more than the largest creditor. It is nearly always the residual claimant on the failed bank’s assets. As the residual claimant, it has the incentive to make socially optimal decisions. The FDIC typically accounts for the vast majority of the claims against the insolvent banks.22 Claims senior to those of the FDIC are nearly always paid in full,23 while the FDIC almost always loses at least some money.24 There are few uninsured deposits that share with the FDIC on a pari-passu basis,25 and claims junior to the FDIC almost never receive a distribution.26 Thus, in most cases the FDIC has the financial incentive to sell the assets of the failed bank for their greatest value. Unlike corporate bankruptcy, bank resolution procedures concentrate decision-making in a single entity with the financial interest in making the right decision about how to dispose of the assets.

Although we argue in favor of the concentration of control in the FDIC, the case has four major caveats. First, a firm’s capital structure depends at least in part in the mechanism used to resolve failed firms; the lack of claims junior to the FDIC may reflect the lack of voice given to these claims in a bank resolution. Second, the FDIC does not make decisions, its employees and directors do. As a government agency, the FDIC may not effectively incentivize its employees to make wealth maximizing decisions. Despite this limitation, however, FDIC control may be the best available alternative given the current capital structure of most failed banks. In a very real sense, failed banks have already been nationalized: Applicable priority rules assure that the FDIC almost always owns a failed bank’s assets. As a result, there are few private parties with any real stake in ensuring that the resolution process is well-run, and those that do exist are poorly placed to supervise the process.

Third, although the FDIC appears to have been the residual claimant in nearly all of the bank insolvencies, it may not have been the residual claimant for most of the assets that have passed through the bank insolvency process.

failed banks would simplify procedures. After all, it is in the best interest of the creditor to obtain the maximum recovery as quickly as possible."). 22. See infra Table 1 (displaying the liability structure of failed banks). 23. See infra note 146 and accompanying text (discussing FDIC losses). 24. See infra Figure 2 and accompanying text (estimating FDIC losses as percent of deposits). 25. See infra Table 1 (displaying the liability structure of failed banks). 26. See infra note 147 and accompanying text (noting the rarity with which general claimants receive payments).

992 67 WASH. & LEE L. REV. 985 (2010)

The capital structure of our very largest banks makes it less likely that the FDIC is the residual claimant, and the largest banks are dramatically bigger than those that typically fail. A single failed bank, Washington Mutual, accounts for more than 70% of all of the assets of banks that failed between January of 1995 and the end of May of 2009,27 and the FDIC suffered no losses from its closure.28 It is dangerous to draw conclusions from a sample of one, but there are other reasons to believe that the FDIC would not be the residual claimant if other extremely large banks failed. The law could adopt special procedures for these banks that force the FDIC to share control. However, doing so could generate substantial costs of their own, principally in administering a special resolution scheme and in the incentives that it would create. Finally, a number of scholars have suggested "automatic" debt conversion methods designed to eliminate the need for formal bank resolution procedures. This Article takes the capital structure of banks as given and therefore does not address proposals that require changes in capital structure such as the issuance of convertible debt. We also spend little time on proposals that would apply without a change in capital structure because these resolution methods will not complete the resolution process. Debt conversion would leave the FDIC as the dominant shareholder of nearly all failed banks. The FDIC is unlikely to want to operate the failed bank going forward and would need to dispose of the assets.

This Article is structured as follows. Section II contrasts bankruptcy and the bank resolution process. Assessments of the existing bank resolution system must take into account the capital structure of failed banks, particularly their liability structure and loss to the FDIC in resolving failed banks. Accordingly, Section III uses data from past bank failures to present the best case for the existing bank resolution system. The best case, it concludes, is a strong one. Section IV describes caveats that limit but do not undermine that case. Section V applies these arguments to recent legislation that extends the FDIC’s control to certain insolvent bank holding companies. Section VI concludes.

27. According to the FDIC, Washington Mutual Bank had assets of $307 billion; the total value of assets of banks that failed between January 1, 1995 and October 19, 2009 was $417 billion. The percentage of assets held by failed "mega-banks" rises considerably if one includes the bank subsidiaries of Citigroup and Bank of America. However, these banks received open-bank assistance and did not formally fail. 28. See infra note 179 and accompanying text (noting that the FDIC transferred most of the assets to J.P. Morgan Chase bank as part of a purchase and assumption transaction).


II. Bankruptcy and Bank Insolvency

Most banks (and nearly all big banks) are owned by bank holding companies.29 Bank holding companies can, and do, file for bankruptcy, but their bank subsidiaries are ineligible for bankruptcy relief.30 For example, Washington Mutual, Inc. filed for bankruptcy,31 but its subsidiary, Washington Mutual Bank, entered receivership and its resolution was governed by a very different set of laws. This is more than just a matter of the organization of statutes, administrative convenience, or discrete technical rules. A more fundamental contrast concerns control: Unlike decision-making under any of the Chapters of the Bankruptcy Code, the law of bank resolution gives decision-making power to one party, the FDIC, and the FDIC can exercise this power with almost no judicial review.

A. Dispersed Control in Bankruptcy

We argue that the primary difference between bankruptcy and bank insolvency is the distribution of control. We therefore must describe the distribution of control in bankruptcy. Control begins with the initiation of the process, and the initiation decision is very different for a bank than it is for other firms. Only a bank’s primary regulator,32 and the FDIC in some cases,33 can place it in receivership. By contrast, a nonbank firm can

29. See Bank Holding Companies and Financial Holding Companies, available at http://www.fedpartnership.gov/bank-life-cycle/charts/bank-ownership-by-bhcs.pdf (showing that bank holding companies own over 83% of banks in the United States and almost all banks with assets of at least one billion dollars). 30. See 11 U.S.C. § 109(b)(2), (d) (2006) (defining when a person may be a debtor under chapter 7 of this title). 31. See 10 Largest Bankruptcies in the History of the USA, NYSE-TRADE.COM, Aug. 16, 2009, http://nyse-trade.com/httpnyse-trade-ru10-krupnejshix-bankrotstv-v-istorii-ssha/ (last visited Sept. 30, 2010) (finding that Washington Mutual, Inc. is the second largest bankruptcy on record, behind only Lehman Brothers Holdings, Inc.) (on file with the Washington and Lee Law Review). 32. See, e.g., 12 U.S.C. § 1464(d)(2)(A) (2006) ("The Director of the Office of Thrift Supervision may appoint a conservator or receiver for any insured savings association if the Director determines, in the Director’s discretion, that 1 or more of the grounds specified in section 11(c)(5) of the Federal Deposit Insurance Act . . . exists."). The primary regulator is the entity that has issued the bank’s charter. 33. See id. § 1821(c)(4) (stating that in defined circumstances "the Corporation may appoint itself as sole conservator or receiver of any insured State depository institution"); id. § 1821(c)(10) (stating that under certain circumstances the "Board of Directors may appoint the Corporation as sole conservator or receiver of an insured depository institution").

994 67 WASH. & LEE L. REV. 985 (2010)

voluntarily file for bankruptcy,34 or a coalition of its creditors can force it into bankruptcy if it is not paying its debts.35 For this reason, creditors have some control over the initiation of a bankruptcy case; they have almost no control in the initiation of the bank resolution process.36 A number of scholars have examined control over the initiation decision.37 We focus instead on control in the resolution process—on control once the process has started.

Control is the power to decide what to do with the firm’s assets. It is a continuous variable, not a discrete one. Accordingly, claimants can have more or less control over that decision. In bankruptcy, stakeholders have different degrees of control over asset sales, depending on the bankruptcy Chapter under which the firm seeks relief. In Chapter 7, for example, the firm’s assets are liquidated by a trustee who can in turn be chosen by the firm’s unsecured creditors.38 The trustee decides how best to liquidate the assets, but her decision is subject to judicial approval,39 and other stakeholders can appear at the hearing and ask the court to stop the sale. The presence of judicial oversight gives the trustee (and the unsecured creditors) less than full control in the matter.

Most firms with asset sizes of even the smallest banks file in Chapter 11.40 In Chapter 11, the managers of the firm retain control as the "debtor in 34. See 11 U.S.C. § 301 ("A voluntary case under a chapter of this title is commenced by the filing with the bankruptcy court of a petition under such chapter by an entity that may be a debtor under such chapter."). 35. See id. § 303 ("An involuntary case against a person is commenced by the filing with the bankruptcy court of a petition . . . by three or more entities, each of which is either a holder of a claim against such person or an indenture trustee representing such a holder . . . ."). 36. See supra notes 32–35 and accompanying text (contrasting bankruptcy and receivership procedures). Even ignoring involuntary bankruptcy, creditors can partially control initiation by controlling access to credit or collateral. In theory, creditors could influence the regulator’s decision as to when to seize the bank. On the other hand, we show below that nondeposit creditors play an insignificant role with most banks. See infra notes 121–23 (discussing the position of nondeposit creditors). 37. See supra note 18 (citing articles that examine initiation issues). 38. See 11 U.S.C. § 702 ("A creditor may vote for a candidate for trustee only if such creditor holds an allowable, undisputed, fixed, liquidated, unsecured claim of a kind entitled to distribution . . . ."). 39. See id. § 363(b) (requiring judicial approval for sales outside the ordinary course of business). 40. See Ed Flynn, Gordon Bermant Burke & Suzanne Hazard, Bankruptcy by the Numbers, JUSTICE.GOV, http://www.justice.gov/ust/eo/public_affairs/articles/docs/abi1220 02.htm (last visited Sept. 27, 2010) (stating that the total amount distributed from all Chapter 7 cases in 2002 was $1.45 billion, and only 1.2% of Chapter 7 cases had more than $500,000 in assets) (on file with the Washington and Lee Law Review). The very smallest bank insolvency in our sample (Monument National Bank) had more than fourteen times this


possession." As such, they exercise the powers of the trustee.41 If the firm is to emerge from bankruptcy, the managers must win confirmation of a plan of reorganization.42 This plan will answer three fundamental questions. First, how will the assets of the firm be used? Second, what is the value of the firm’s assets? Third, how will the proceeds of the assets be divided amongst the various claimants?

There are two ways for a plan to be confirmed. One is by the consent of creditors.43 The other is by judicial confirmation of the plan under prescribed conditions over the objection of creditors.44 Creditor consent is determined by complicated voting procedures that can grant substantial power to minority creditors. Claims and interests are divided into classes, and half of the claims cast (two-thirds by value) of each class must vote in favor of the plan.45 Dissimilar claims, such as secured and unsecured claims, cannot be placed in the same class,46 and consensual confirmations require unanimity among classes.47 Thus, a group of creditors or even equity holders can block confirmation if they believe that they are entitled to or can demand more. Junior claimants can sometimes demand more simply because the alternative to a consensual plan is so difficult and costly. Most significantly, the court must determine that the plan achieves both horizontal equity (the plan does not unreasonably discriminate between creditors of equal priority) and vertical equity (the plan does not pay junior creditors anything if senior

amount in assets ($7,680,000). 41. See 11 U.S.C. § 1107(a) (requiring judicial approval for sales outside the ordinary course of business). 42. This assumes that the plan will be confirmed during the period in which only the debtor can propose a plan. 43. See 11 U.S.C. § 1129(a) (stating that a court will only confirm a plan if "each class of claims or interest . . . has accepted the plan"). 44. See id. § 1129(b) (stating that the court will confirm the plan without the creditors’ consent if "the plan does not discriminate unfairly, and is fair and equitable with respect to each class or interest that is impaired under, and has not accepted, the plan"). 45. See id. § 1126(c) ("A class of claims has accepted a plan if such plan has been accepted by creditors . . . that hold at least two-thirds in amount and more than one-half in number of the allowed claims of such class held by creditors . . . that have accepted or rejected such plan."). 46. See id. § 1122(a) ("Except as provided in subsection (b) of this section, a plan may place a claim or an interest in a particular class only if such claim or interest is substantially similar to the other claims or interests of such class."). 47. See id. § 1129(a)(8) ("The court shall confirm a plan only if . . . with respect to each class of claims or interests such class has accepted the plan; or such class is not impaired under the plan.").

996 67 WASH. & LEE L. REV. 985 (2010)

creditors are not paid in full).48 Because plans will rarely pay everyone in cash, these findings require complex and time-consuming fact-finding. For this reason, most confirmed reorganization plans are consensual.49

Both consensual and nonconsensual confirmations take time, and the Code authorizes management to make day-to-day operating decisions prior to confirmation.50 The Code also allows management to make some decisions, with court approval, that are not in the ordinary course of the debtor’s business.51 Among the most important is management’s decision, usually at the urging of a creditor, to sell some or most of the debtor’s assets. Bankruptcy practitioners have found that they can use this power to determine how the assets will be used without having to seek plan approval. Chapter 11 sales of going concerns outside a reorganization plan, while once rare, have now become frequent.52 48. See id. § 1129(b) (requiring a plan to be fair and equitable). 49. See supra note 4 and accompanying text (recognizing that an overwhelming majority of reorganizations are resolved consensually with the approval of each class of creditors and shareholders). 50. 11 U.S.C. § 363. 51. Id. § 363(b)(1). 52. See Douglas G. Baird, The New Face of Chapter 11, 12 AM. BANKR. INST. L. REV. 69, 71 (2004) ("Today, creditors of insolvent businesses . . . no longer need a substitute for a market sale. Instead of providing a substitute for a market sale, chapter 11 now serves as the forum where such sales are conducted."); Douglas G. Baird & Robert K. Rasmussen, Chapter 11 at Twilight, 56 STAN. L. REV. 673, 674 (2003) ("Hence, modern Chapter 11 practice cannot be squared with the traditional account. Regardless of whether the number of businesses entering Chapter 11 rises or falls, something different is going on."); Douglas G. Baird & Robert K. Rasmussen, The End of Bankruptcy, 55 STAN. L. REV. 751, 751 (2002) ("Corporate reorganizations have all but disappeared. Giant corporations make headlines when they file for Chapter 11, but they are no longer using it to rescue a firm from imminent failure. Many use Chapter 11 merely to sell their assets and divide up the proceeds."); see generally George W. Kuney, Hijacking Chapter 11, 21 EMORY BANKR. DEV. J. 19 (2004). According to the literature, Chapter 11 sales of going concerns outside a reorganization plan were unusual between 1978 and the mid-1990s. By 2000, such sales were frequent, not the exception. In 2002, for instance, a majority of large firms in Chapter 11 were sold in one form or another. Note that the classification of asset dispositions into reorganizations or asset sales is somewhat subjective. Nearly every bankrupt debtor in Chapter 11 reorganization will use § 363 to sell at least some assets. Section 363(b) authorizes the sale of assets in the ordinary course without court approval. The trustee, acting as the debtor in possession, often will sell assets to maintain liquidity or avoid losses on assets with declining values. We are more interested in sales of substantial portions of the firm outside of the ordinary course of business. By one count, about 16% were auctioned by means of a 363 sale. See LYNN M. LOPUCKI, COURTING JUSTICE 170–71 (2005) (providing sale distributions). The trend continues to be prevalent to date. See Mike Specter, GM Asset Sale Gets Judge’s Nod, WALL ST. J., July 6, 2009, at B1 (providing a summary of the GM sale); see also CONGRESSIONAL OVERSIGHT PANEL, supra note 10 (providing a summary of the details of sale of both Chrysler and GM).


The recent automobile bankruptcies provide salient examples of this trend. The Chrysler and General Motors groups filed for bankruptcy protection under Chapter 11. Rather than seek approval of their creditors to a plan of reorganization, each quickly sold its most valuable assets to newly formed entities in exchange for cash and the assumption of some of the old firms’ liabilities.53 The old corporate groups remained behind in bankruptcy with the unwanted assets and the disfavored creditors, but the press reported that the firms emerged from bankruptcy. As a practical matter the press was correct: The firm’s critical assets and operations emerged from bankruptcy; only the disfavored creditors and assets retained by the old corporate shells remained behind.

This description implies that management controls the process of disposing of assets. As a formal matter it does. As a practical matter, however, creditors often decide how the firm’s assets will be disposed because they enjoy sufficient leverage to effectively dictate the management’s decisions.54 Once again, the automobile bankruptcies provide a good example. The continued existence of Chrysler and General Motors (and thus the fate of the senior management) largely depended on continued funding. Only the federal government was willing to lend. Management was understandably loath to make decisions that would anger the administration. Similarly, management of other bankrupt firms may be unwilling to make decisions that would anger their dominant creditors, particularly if these creditors included strong covenants in their loan agreements. Management jeopardizes its continued employment with the firm and perhaps future employment prospects by making decisions that creditors with leverage dislike.55 However, neither management nor a dominant creditor will have

53. See Mike Specter, GM Asset Sale Gets Judge’s Nod, WALL ST. J., July 6, 2009, at B1 (providing a summary of the GM sale); see also CONGRESSIONAL OVERSIGHT PANEL, supra note 10 (providing a summary of the details of sale of both Chrysler and GM). 54. See Douglas G. Baird, The New Face of Chapter 11, 12 AM. BANKR. L. REV. 69, 75 (2004) ("The ability of creditors to control their debtor and negotiate with each other outside of chapter 11 is now vastly greater than it was during the equity receivership—or even in chapter 11 just 20 years ago."); Barry E. Adler et al., Destruction of Value in the New Era of Chapter 11 1 (N.Y.U. Stern Sch. of Bus. Research Paper Series, Working Paper No. CLB-06-032, 2006), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id= 1291620 (noting that it is common for creditors to control the bankruptcy process); Kenneth M. Ayotte & Edward R. Morrison, Creditor Control and Conflict in Chapter 11 1 (Columbia Univ. Ctr. for Law & Econ., Research Paper No. 321 and Northwestern Univ. Law Sch. Law & Econ. Research Paper Series, No. 08-16 (2008), available at http://papers.ssrn.com/ sol3/papers.cfm?abstract_id=1081661 (finding pervasive creditor control in the bankruptcy process). 55. See M. Todd Henderson, Paying CEOs in Bankruptcy: Executive Compensation

998 67 WASH. & LEE L. REV. 985 (2010)

complete control. By selling assets outside of the plan of reorganization, they successfully deprive other claimants of the right to vote to block the sale. But they must still seek judicial approval of the sale, and these other claimants have the right to argue that the sale is not in the best interest of the claimants as a group. Management, on the sale proponents’ behalf, must provide a business reason for the sale, and it must convince the court that the sale maximizes the value of the estate. Thus, in an important sense, management, dominant creditors and the bankruptcy court share control over disposal of the firm’s assets. As we note below, a single creditor, the FDIC, enjoys complete control over bank resolutions. The sale of a bank’s assets does not require judicial approval.

B. Bank Insolvencies and the Purchase and Assumption Agreement

Bank receiverships and bankruptcy proceedings operate under significantly different rules.56 The differences include the procedure for determining claims, the right to repudiate contracts, stays of litigation, and the power to avoid certain transactions. Claims against a nonbank debtor are allowed or disallowed by the bankruptcy court, and its determination can be appealed.57 By contrast, the authority to disallow claims is given to the FDIC, as receiver; its determination is subject to limited judicial review.58 The FDIC has to power to repudiate or perform contracts entered into by the failed bank;59 the bankruptcy trustee can reject or assume only contracts that

When Agency Costs Are Low, 101 NW. U. L. REV. 1543, 1595–96 (2007) (discussing how Chapter 11 is no longer "an entrenchment mechanism for managers that allows them to stay in power and extract rents from the firm"); Stuart C. Gilson & Michael R. Vetsuypens, CEO Compensation in Financially Distressed Firms: An Empirical Analysis, 48 J. FIN. 425, 456 (1993) (reviewing management compensation policies for firms that have filed for bankruptcy and finding that nearly one-third of all CEOs are being replaced). 56. American banks have been barred from bankruptcy courts since the inception of the first lasting bankruptcy act in 1898. See Bankruptcy Act of 1898, ch. 541 § 4(b), 30 Stat. 544, 547 (explaining "who may become bankrupt" under the Act). Prior to 1933, state receivership law controlled state bank failures, and federal law governed national bank failures. Today, the Federal Deposit Insurance Act governs failures of FDIC-member banks. 57. See 11 U.S.C. § 502 (2006) ("[I]f such objection to a claim is made, the court, after notice and a hearing shall determine the amount of such claim . . . ."). 58. See 12 U.S.C. § 1821(d)(5)(E) (2006) ("No court may review the Corporation’s determination pursuant to subparagraph (D) to disallow a claim."). Courts rarely disagree with the FDIC’s disallowance of a claim. See, e.g., Adagio Inv. Holdings Ltd. v. FDIC, 338 F. Supp. 2d 71, 79 (D.D.C. 2004) ("FDIC-R makes much of its broad power as a receiver, but these powers are not without limits . . . ."). 59. See 12 U.S.C. § 1821(e)(1) ("In addition to any other rights a conservator or


are executory.60 Litigation against the failed bank is not automatically stayed while the bank’s failure is being resolved. The FDIC must submit a timely request asking a court to enjoin litigation;61 and courts disagree as to whether they must comply with the request.62 Applicable law does not otherwise prevent creditors from enforcing their property rights in the failed bank’s assets.63 Bankruptcy’s automatic stay prevents the enforcement of property interests (such as security interests) in the assets of the debtor.64 Finally, certain payments made by nonbank firms to their creditors are recoverable by the firms’ trustees as preferences.65 Depositors are creditors of their depository banks. However, withdrawals by depositors before the bank fails are not recoverable by the FDIC as preferences. The FDIC’s power to avoid fraudulent transfers by the failed bank is more limited than the comparable power given to the bankruptcy trustee. Unlike the bankruptcy trustee’s power,66 the FDIC’s avoidance power requires actual fraud;67 constructive fraud is insufficient. These are important difference between bank and bankruptcy insolvency rules. However, we argue that the most critical

receiver may have, the conservator or receiver for any insured depository institution may disaffirm or repudiate any contract or lease . . . ."). 60. See 11 U.S.C. § 365(a) ("[T]he trustee, subject to the court’s approval, may assume or reject any executory contract or unexpired lease of the debtor."). 61. See 12 U.S.C. § 1821(d)(12)(A) ("After the appointment of a conservator or receiver for an insured depository institution, the conservator or receiver may request a stay for a period not to exceed 45 days, in the case of a conservator; and 90 days in the case of any receiver . . . ."). 62. Compare Wachovia Bank, N.A. v. Michael Taylor, 2007 U.S. Dist. LEXIS 83793, at *5 (W.D. Mich. Nov. 13, 2007) (finding that "a stay is mandatory and must be granted"), with FDIC v. Taylor, 727 F. Supp. 326, 328 (S.D. Tex. 1989) (finding that the stay is permissive because "[t]he statute was not enacted to give FDIC the power to stay proceedings to which it is a party at any point, regardless of the length of its involvement."). 63. See FDIC Advisory Opinion: Self-Help Liquidation of Collateral by Second Claimants in Insured Depository Receiverships (Dec. 15, 1989), FDIC 89-49 ("[I]t is my opinion that such a secured creditor of an insured depository institution for which a receiver had been appointed could liquidate the creditor’s properly pledged collateral by commercially reasonable ‘self-help’ methods, provided that no involvement of the receiver was required . . . ."). 64. See 11 U.S.C. § 362 (stating that a petition operations as a stay of "the enforcement, against the debtor or against the property of the estate, of a judgment obtained before the commencement of the case under this title"). 65. See id. § 547 (discussing when transferees may or may not avoid a transfer). 66. See id. § 548(a)(1) (outlining the bankruptcy trustee’s avoidance power). 67. See 12 U.S.C. § 1821(d)(17) (2006) (stating that the FDIC may avoid a transfer or incurrence of any liability only if such action was made with "the intent to hinder, delay, or defraud").

1000 67 WASH. & LEE L. REV. 985 (2010)

differences are the allocation of control over the resolution process and the distribution of losses among claimants.

Somewhat different priority rules order claims against banks and nonbanks. Regardless of whether the firm is a bank or nonbank, secured claims have first priority with respect to their collateral,68 administrative expenses should be paid before unsecured creditors,69 and subordinated debt and equity should receive nothing unless general unsecured creditors are paid in full. The priority rules differ, however, in the extent by which they distinguish between unsecured creditors. Bankruptcy does grant some unsecured creditors priority over others; for example, workers and customers are given limited priority over general claimants.70 In large corporate bankruptcies these priority claims do not typically account for a substantial portion of the total claims.71 By contrast, banking law grants priority to domestic deposits over foreign deposits and general claims.72 The FDIC insures domestic deposits up to an amount that varies over time, currently $250,000,73 and becomes subrogated to the claims of these creditors.

68. See id. § 1821(11)(A) (outlining the order of priority for collateral); 11 U.S.C. § 725 ("After the commencement of a case under this chapter . . . the trustee . . . shall dispose of any property in which an entity other than the estate has an interest, such as a lien, and that has not been disposed of under another section of this title."). 69. See 12 U.S.C. § 1821(d)(11)(A)(i) (stating that the "administrative expenses of the receiver" have priority over unsecured creditors); 11 U.S.C. § 507(a)(2) (placing administrative expenses before unsecured claims in the order of priority). 70. See 11 U.S.C. § 507(a)(4) ("Fourth, allowed unsecured claims, but only to the extent of $10,000 for each individual or corporation . . . ."); id. § 507(a)(7) ("Seventh, allowed unsecured claims of individuals, to the extent of $1,800 for each such individual . . . ."). 71. See Elizabeth Warren & Jay Westbrook, Contracting Out of Bankruptcy, 118 HARV. L. REV. 1197, 1243 (2004) (showing that employee wage claims account for 8.8% of the debt of business debtors in 1994). Bankrupt debtors do, however, sometimes repay "critical vendors" at the outset of the case even though other creditors will not be repaid in full. However, the authority to do so is unclear. See In re Kmart Corp., 359 F.3d 866, 874 (7th Cir. 2004) ("Even if § 362(b)(1) allows critical-vendors in principle, preferential payments to a class of creditors are proper only if the record shows the prospect of benefit to the other creditors."). 72. See 12 U.S.C. § 1813(l)(5) (stating that "any deposit liability of the institution" receives preference over "any other general or senior liability of the institution"). 73. See Dodd-Frank Wall Street Reform and Consumer Protection (Dodd-Frank) Act, Pub. L. No. 111-203, §§ 201(a)(11), 101(a)(1) (West, Westlaw through 2010 legislation), H.R. 4173, 111th Cong. §§ 201(a)(11), 202(a)(1) (2010) (to be codified at 12 U.S.C. § 5323) (defining the term "financial company" and instructing how to properly commence an orderly liquidation); 12 U.S.C. § 1821(a)(1)(A) (2006) ("The Corporation shall insure the deposits of all insured depository institutions as provided in this chapter."). The insurance limit applies to the depositor, not the deposit. See id. § 1813(m)(1) ("[T]he term ‘insured


Figure 1 presents the FDIC’s June 2009 categorization of how it dealt with troubled banks between January of 1995 and May of 2009.74 The FDIC can provide assistance without closing the bank (the FDIC calls these "Assistance Transactions" or "A/A").75 This includes purchasing nonvoting securities in the bank or assuming some of its liabilities.76 Its purpose is to deposit’ means the net amount due to any depositor for deposits in an insured depository institution as determined under sections 1817(i) and 1821(a) of this title."). 74. In generating Figure 1, we calculate each bank as a separate transaction even if the banks are part of a related family. For example, the assistance that the FDIC provided to Citigroup and its affiliates counted as five transactions because there were five banks. All figures and tables presented in this Article reflect the data we gathered in the summer of 2009. The FDIC occasionally revises this data. For example, Figure 1 does not include the assistance provided to Bank of America and its affiliates in January of 2009 because these transactions were not added to the FDIC database until after June of 2009. 75. See 12 U.S.C. § 1823(c)(1) ("The Corporation is authorized, in its sole discretion and upon such terms and conditions as the Board of Directors may prescribe, to make loans to, to make deposits in, to purchase the assets or securities of, to assume the liabilities of, or to make contributions to, any insured depository institution . . . ."); id. § 1823(c)(3) ("The Corporation may provide any person acquiring control of, merging with, consolidating with or acquiring the assets of an insured depository institution under subsection (f) or (k) of this section with such financial assistance as it could provide an insured institution under this subsection."). Another open bank transaction is a reprivatization in which management takes over the bank and sells it with or without the assistance of the FDIC. We do not discuss this transaction as it has not been used since 1989 and has been used just three times since 1934. 76. Id. § 1821(a)(3), (c)(5).

A/A4%

IDT2%

PO6%

PI34%

PA54%

Figure 1: Failure and Assistance Transactions from 1/1/2005 to

5/31/2009

1002 67 WASH. & LEE L. REV. 985 (2010)

reinforce the capital of a troubled bank, making the bank more attractive to investors, and avoid having put the bank through the resolution process. Assistance transactions are rare, however. The only assistance transactions that occurred in our sample period were the FDIC’s assistance to the five bank subsidiaries of Citigroup77 and the eight bank subsidiaries of Bank of America.78

If the FDIC decides to close a failed bank, it has several options. It can simply pay the insured depositors what they are owed (a "pay-out" or PO) or transfer the insured deposits to another bank (an "insured deposit transfer" or IDT). In both cases, the FDIC liquidates the assets of the bank and distributes the proceeds in accordance with the priority structure. Like open-bank assistance, these choices of resolution are rare, occurring only about 8% (6% PO and 2% IDT) of the time since 1995.

In 88% of cases the FDIC found a bank that was willing to assume some or all of the failed bank’s liabilities and purchase some or all of the failed bank’s assets. In a significant portion of the cases (34%), the acquirer assumed only the insured liabilities (a purchase and assumption of insured deposits, or "PI"). However, most of the time (54%), the acquirer assumed some of the uninsured deposits as well (a purchase and assumption agreement or "PA"). These transactions also vary in the nature of the assets acquired. The standard purchase and acquisition contract grants the acquirer the option to purchase the physical assets of the bank (the offices, furniture, etc.),79 but on average these assets account for only about 2.3% of total assets; the

77. The FDIC and the U.S. Treasury guaranteed $306 billion of loans and securities held by these banks. Under a loss sharing arrangement both agreed to bear part of the banks’ losses on these assets above $29 billion, up to stipulated amounts. The arrangement also gave the FDIC and Treasury preferred shares in Citicorp’s banks. See Press Release, Joint Statement by Treasury, Federal Reserve and FDIC on Citigroup (Nov. 23, 2008), http://www.federalreserve.gov/newsevents/press/bcreg/20081123a.htm (last visited Sept. 28, 2010) ("As a fee for this arrangement, Citigroup will issue preferred shares to the Treasury and FDIC.") (on file with the Washington and Lee Law Review). 78. The FDIC and the U.S. Treasury guaranteed the value of $118 billion of assets held by Bank of America and its subsidiaries. In return the FDIC and the Treasury received preferred shares in Bank of America. See Press Release, Treasury, Federal Reserve and the FDIC Provide Assistance to Bank of America (Jan. 16, 2009), http://www.fdic.gov/ news/news/press/2009/pr09004.html (last visited Sept. 28, 2010) ("Treasury and the Federal Deposit Insurance Corporation will provide protection against the possibility of unusually large losses on an asset pool of approximately $118 billion of loans, securities backed by residential and commercial real estate loans, and other such assets . . . .") (on file with the Washington and Lee Law Review). 79. See generally FDIC, PURCHASE ASSUMPTION AGREEMENT (2009), available at http://www.fdic.gov/bank/individual/failed/sb_P_and_A.pdf (providing a sample purchase and assumption agreement).


primary assets of the bank are its cash, loans and securities.80 In most cases (51%) the acquirer acquired less than one-quarter of the assets of the failed bank;81 the remaining assets were sold by the FDIC in the ensuing months or even years.82 When the acquirer acquired the majority of the assets of the failed bank, the FDIC usually entered into a loss-sharing agreement so that it retained much of the risk of a decline in value. The typical arrangement divides loss around specific acquired assets by threshold amounts or "tranches." In the first tranche the acquirer bears all loss up to a predetermined threshold amount. Loss in the second tranche is shared between the FDIC and acquirer, with the FDIC bearing most of it. The FDIC bears almost all of the loss in the third tranche.83 Of the twenty-nine insolvencies in which the acquirer purchased more than three-quarters of the firm’s assets, the FDIC entered loss-sharing agreements in eighteen of the transactions. This ratio rises to sixteen of nineteen transactions in the banking crisis of 2008 and 2009. The FDIC entered into just one loss-sharing transaction in the sixty-nine transactions in which the acquirer took less than three-quarters of the failed bank’s assets.

Since 1991, Congress has required the FDIC to choose the resolution method that imposes the least cost on the insurance fund, unless the FDIC determines that doing so is necessary to avert systemic risk.84 However, in order to invoke this exception to the least cost rule the FDIC must obtain the approval of the Chairman of the Federal Reserve and the Treasury Secretary, and they must consult with the President.85 This exception has been invoked

80. See FREDERICK S. MISHKIN, THE ECONOMICS OF MONEY, BANKING AND FINANCIAL MARKETS 226 (9th ed. 2010) (illustrating that in 2008, nine percent of total commercial bank assets were "other assets," which include physical capital). 81. In about 51% of the Purchase and Assumption (PA, PI and P&A) transactions for which the FDIC provided data on the assets acquired the acquiring bank purchased less than 25% of the failed bank’s assets. In another 12% of these transactions the acquirer purchased between 25% and 50% of the assets. The remaining transactions were between 50% and 75% (7% of transactions) and 75% and 100% (30% of transactions). 82. See supra note 13 and accompanying text (providing a timeline for the disposition of a failed bank’s assets). 83. See, e.g., Purchase and Assumption Agreement of FDIC and Wilshire State Bank, June 26, 2009, § 4.15, available at http://www.fdic.gov/bank/individual/failed/mirae_ P_and_A.pdf (outlining the agreement with respect to loss sharing between the parties). 84. See 12 U.S.C. § 1823(c)(4)(A) (2006) (providing the limited circumstances under which the least-cost resolution method will not be required). 85. See id. § 1823(c)(4)(G) (stating that "the Corporation may take other action or provide assistance under this section as necessary to avoid or mitigate such effects" with the approval of the Board of Directors, the Board of Governors of the Federal Reserve System, and the Secretary of the Treasury).

1004 67 WASH. & LEE L. REV. 985 (2010)

just twice since 1995 in order to provide assistance to Citigroup, Bank of America and their subsidiaries.86 Prior to the adoption of the least cost standard, the FDIC used standards that allowed them to consider, among other things, the impact of a bank closure on a community.87 Congress worried that the FDIC was too willing to consider these other factors and too willing to pay an acquiring bank to assume all of the liabilities of the failed firm, not just the insured liabilities. The effect of such payments was to use the federal deposit insurance fund to subsidize payments to uninsured depositors and other creditors because the FDIC typically would have to pay an acquiring bank more to assume all of a failed bank’s deposits than assume only its insured deposits. Accordingly, some observers believed that resolutions involving the assumption of all deposits would become the exception rather than the norm as they would not satisfy the least cost requirement.88 Data on the form of resolution used by the FDIC between 1995 and 2009 clearly show that this has not been the case as transactions in which the acquirer assumes some uninsured deposits outnumber those in which they do not by a factor of about two to one.89

Some purchase and assumption transactions can be quite similar to bankruptcies that utilize Section 363 to sell all or substantially all of the assets. Both processes can be used to quickly transfer the core assets or goodwill of the failed entity to an acquirer. Assets not transferred are retained for months or even years before being liquidated and creditors paid in accordance with the relevant priority rules.90 However, the processes 86. The FDIC also invoked this exception when approving financing for Citibank’s bid to buy Wachovia. See Editorial, Who’s Too Big to Fail?, WALL ST. J., Sept. 13, 2009, at A14 ("To provide assistance, the [FDIC] board had to invoke the ‘systemic risk’ exception in the Federal Deposit insurance Act . . . ."). However, this transaction was not consummated as Wells Fargo purchased Wachovia instead. See id. ("Yet days later, Wachovia cut a better deal to sell itself to Wells Fargo, instead of Citi."). 87. Prior to 1991, the Federal Deposit Insurance Act required the FDIC to dispose of a failed bank’s assets in such a way that the cost to the insurance fund was less than the cost to it of paying off insured deposits. 12 U.S.C. § 1823(c)(4)(A). However, 12 U.S.C § 1823(c)(4)(A) recognized an exception allowing the FDIC to use a method of disposal not satisfying this "less cost" standard if the failed bank was deemed essential to its community. Id. 88. See Skeel, supra note 18, at 771 ("As a result, the current framework should curtail significantly regulators’ ability to favor purchase and assumption transactions over liquidation and insured deposit transfers."). 89. See supra Figure 1 (graphing failure and assistance transactions from 2005 to 2009). The dominance of PA transactions is even more apparent if we expand our sample to include all transactions from 1991 through May of 2009. During this period there were 135 PI transactions and 513 PA or P&A transactions. 90. Lehman Brothers Holdings, Inc.’s bankruptcy shows the same pattern. See


differ in three fundamental ways. First, the initial stage of the bank resolution process is much faster than even the quickest bankruptcy. Chrysler and General Motors were notable because the sales were completed in a matter of weeks, and the fastest 363 sales take more than a week.91 In every completed purchase and assumption transaction, the FDIC completed the sale of the failed bank simultaneously with its seizure.

Second, the standards for determining the method of resolutions are not the same. Surprisingly, the Bankruptcy Code does not provide the judge with an explicit standard for approving sales outside of the ordinary course of business.92 The standard has been left to courts, which typically require the sales proponent to articulate a "sound business reason" for the sale.93 Many courts and commentators believe that the judge should approve a transaction if, and only if, it increases the aggregate return of all parties with legal claims against the failed entity.94 Other commentators disagree. They believe that the court should also consider the interests of other stakeholders who lack a legal claim such as workers or the local community.95 Congress explicitly rejected this broader "stakeholder" standard in the context of bank insolvencies when its own money was on the line.96 It did, however, provide a possible exception in the case of systemic risk, but this exception requires an acknowledgement of those who face far more political accountability than

Lehman’s Collapse: A Timeline, DEAL, Sept. 26, 2009, http://www.thedeal.com/dealscape/ 2009/09 (last visited Sept. 30, 2010) (providing a timeline of the collapse of Lehman Brothers Holdings, Inc.) (on file with the Washington and Lee Law Review). The largest bankruptcy in U.S. history, Lehman’s North American investment banking unit and headquarters was sold in a 363 sale within a week of its bankruptcy filing. Id. Other core units were sold within weeks later. Still other assets were auctioned approximately three months after the filing. Id. 91. LYNN M. LOPUCKI, COURTING FAILURE: HOW COMPETITION FOR BIG CASES IS CORRUPTING THE BANKRUPTCY COURTS 171–72 (2005). 92. See 11 U.S.C. § 363 (2006) (governing the use of sale or lease in the ordinary course of business). 93. See Comm. of Equity Sec. Holders v. Lionel Corp. (In re Lionel Corp.), 722 F.2d 1063, 1070 (2d Cir. 1983) ("The rule we adopt requires that a judge determining a § 363(b) application expressly find from the evidence presented before him at the hearing a good business reason to grant such an application."). 94. See, e.g., Douglas G. Baird, Bankruptcy’s Uncontested Axioms, 108 YALE L.J. 573, 576 (1998) (identifying and contrasting two separate approaches of bankruptcy experts); Douglas G. Baird, Loss Distribution, Forum Shopping, and Bankruptcy: A Reply to Warren, 54 U. CHI. L. REV. 815, 833 (1987) (arguing the basis of bankruptcy policy in the context of business bankruptcies and defending the "creditors’ bargain" approach). 95. See, e.g., Elizabeth Warren, Bankruptcy Policy, 54 U. CHI. L. REV. 775, 778–93 (1987) (expounding upon the traditional conception of the purposes of bankruptcy). 96. Supra notes 84–85 and accompanying text.

1006 67 WASH. & LEE L. REV. 985 (2010)

a bankruptcy judge.97 Significantly, this exception has only been invoked twice to preserve groups of banks owned by Citigroup and Bank of America.98

The third difference is the most significant. While control of the bankruptcy process is divided among all of the claimants, control over a bank receivership is concentrated in the FDIC. Bankruptcy provides all claimants with the opportunity to vote on the plan and to at least object to a sale outside of the ordinary course of business.99 When objecting to the sale, claimants can, and do, argue that the sale of the assets would yield more if the assets had been marketed in another manner or if an alternative buyer had been chosen. If the judge believes that the dissenters have the better of the argument, she can enjoin the sale. By contrast, claimants on the assets of the failed bank have neither the right to vote on the sale of the key assets of the bank nor the ability to seek an injunction from a court.100 They can, and sometimes do, seek damages if they believe that the FDIC made poor decisions. However, claimants are entitled only to the difference between what they received and the amount that they would have received in liquidation.101 They have no entitlement to any going concern value of the bank.102

III. The Case for FDIC Control of Bank Insolvencies

We argue that the key difference between a bank receivership and a bankruptcy proceeding is the concentration of control in a single decision

97. Supra notes 84–85 and accompanying text. 98. The FDIC also invoked this rule to arrange Citibank’s acquisition of Wachovia, but Wachovia was ultimately sold to Wells Fargo. See Joe Adler, Wachovia Deal Forces FDIC to Use Systemic Tool, AM. BANKER, Sept. 30, 2008, at 4 ("The Federal Deposit Insurance Corp. used the ‘systemic risk’ exception for the first time in history Monday so it could creatively find a way to sell Wachovia Corp.’s assets to Citigroup Inc. while protecting the government from losses."). 99. See supra notes 39–40 and accompanying text (describing the bankruptcy process). 100. See 12 U.S.C. § 1821(d)(13)(D) (2006) (stating no court shall have jurisdiction over "any claim or action for payment from, or any action seeking a determination of rights with respect to, the assets of any depository institution for which the Corporation has been appointed receiver"). 101. See id. § 1821(i)(2) ("The maximum liability . . . shall equal the amount such claimant would have received if the Corporation had liquidated the assets and liabilities of such institution . . . ."). 102. Id.


maker—the FDIC. The FDIC is not a neutral arbitrator, and the concentration of control in its hands can impose costs if it acts against the interests of other creditors.103 The literature offers two plausible justifications for this concentration of control. First, it allows for much greater speed and secrecy, and this speed and secrecy could yield benefits in banking. Second, concentration of control is appropriate to the extent that the FDIC is the largest creditor and will wish to maximize the value of the assets of the firm. We find the first justification unconvincing. However, if we recast the second justification to say that the FDIC is the residual claimant on the assets of most failed banks, it provides a strong rationale for FDIC control.

A. The Benefits of Speed and Secrecy

Numerous commentators ask whether banking regulators act with sufficient speed when seizing failing institutions.104 Banks are highly leveraged, much more so than nonfinancial firms.105 This high leverage means that relatively small shocks can quickly render the firm insolvent. The relatively liquid nature of the bank’s assets means that the moral hazard created by insolvency or near-insolvency can be severe. Banks that are insolvent or close to insolvent can easily deploy their assets quickly to increase risk and exploit the option value of equity in a nearly insolvent firm. The dispersion of depositors makes coordination difficult, and thus initiation rules that differ from those that apply in bankruptcy may be justified. This Article does not, however, address the speed with which the regulator

103. Writing just after the bank crises of 1930–1933, Cyril Upham and Edwin Lamke worried that the recently created FDIC would favor its own interests over those of other creditors. They concluded that "since [the FDIC] is a quasi-governmental body, it may be expected to administer assets fairly with respect to the rights of all depositors." CYRIL B. UPHAM & EDWIN LAMKE, CLOSED AND DISTRESSED BANKS: A STUDY IN PUBLIC ADMINISTRATION 58 (1934). For doubts on this score, see infra Part III.A. 104. See, e.g., EDWARD J. KANE, THE S&L INSURANCE MESS: HOW DID IT HAPPEN? 1 (1985) (arguing the problem rests in structural flaws of the system and that periodic review is imperative); Arnoud W.A. Boot & Anjan V. Thakor, Self-Interested Bank Regulation, 83 AM. ECON. REV. 206, 207–08 (1993) (focusing on bank rescue packages and on the behavior of troubled banks in light of rescue offers); Skeel, supra note 18 (discussing initiation issues proposing possible reforms). 105. Banks have a debt-equity ratio of about 9:1. See generally RISK MGMT. ASS’N, THE ANNUAL STATEMENT STUDIES: FINANCIAL RATIO BENCHMARKS 2008–2009 (2008). Nonbank firms in different industries tend to have much lower leverage ratios. Id. The following are median debt-equity ratios for select industries in 2004–2005: Agriculture (dairy cattle and milk production): 2:3; manufacturing (plastics): 1:4; mining (sand and gravel): 1:4; and transportation (freight trucking): 2:6. Id.

1008 67 WASH. & LEE L. REV. 985 (2010)

initiates the bank resolution process as the bank nears insolvency. Rather, it asks whether there are significant benefits from speed in resolving a failed bank once a procedure has been initiated. We are skeptical of the possible benefits of speed and conclude that speed does not justify giving the FDIC control over the resolution process.

Speed provides two closely related benefits: liquidity and confidence in the banking system. Speed gives depositors uninterrupted access to their deposit balances, preserving liquidity.106 The assurance of continued liquidity of deposits in turn increases depositors’ confidence in the banking system and prevents bank runs that could pose a risk to the entire financial system. Although these benefits may be substantial, they do not require FDIC control over the resolution process.

In resolving a failed bank, the resolving authority must dispose of the bank’s assets and distribute the proceeds to the bank’s creditors. These tasks are distinct from the FDIC’s duty to honor its commitment to the insured depositors. The FDI Act requires the FDIC to reimburse insured depositors from the Deposit Insurance Fund ("DIF" or "insurance fund")107 to the

106. See FDIC, THE FIRST FIFTY YEARS: A HISTORY OF THE FDIC 1933–1983, at 83 (1984) ("[T]here were also conflicting concerns that depositors had to wait too long to recover their funds."). One can question the importance of this liquidity. Depositors can obtain liquid short-term funds by using credit cards or investing in money market funds offered by mutual funds. They do not need the liquidity provided by demand deposits; alternative products offering it are available. See Daniel R. Fischel, Andrew M. Rosenfield & Robert S. Stillman, The Regulation of Banks and Bank Holding Companies, 73 VA. L. REV. 301, 318 (1987) ("This delay in the ability to obtain access to funds has historically been thought to be intolerable in the case of depositors . . . its validity is questionable today because depositors have available liquidity substitutes such as money market funds and credit cards."). For the rise of financial products issued by nonbanks that are functionally similar to demand deposits, see Jonathan Macey & Geoffrey Miller, Nondeposits and the Future of Banking Regulation, 91 MICH. L. REV. 237, 237 (1992). 107. For the history of the Federal Deposit Insurance Act, see Charles W. Calomiris & Eugene N. White, The Origins of Deposit Insurance, in THE REGULATED ECONOMY: A HISTORICAL APPROACH TO POLITICAL ECONOMY 145 (Claudia Goldin & Gary D. Libecap eds., 1994). For the FDIC practice of allowing prompt access, see George G. Kaufman & Steven A. Seelig, Post-Resolution Treatment of Depositors at Failed Banks: Implications for the Severity of Banking Crises, Systemic Risk, and Too Big to Fail, 26 ECON. PERSP. 27, 32 (2002). For the FDIC’s prompt payment of insured deposits when conducting a direct payoff transaction, see George G. Kaufman, Depositor Liquidity and Loss Sharing in Bank Failure Resolutions, 22 CONTEMP. ECON. POL’Y 237, 245 (2004). The United States is one of the few countries that gives depositors immediate access to insured deposits. See Kaufman & Seelig, supra, at 32–33, 36 (noting the countries in Table 1); George G. Kaufman, Using Efficient Bank Insolvency Resolution to Solve the Deposit Insurance Problem, 8 J. BANKING REG. 40, 46 (2006) (noting that FDIC usually pays insured deposits on close of business day after seizure). For the timing in which select countries with deposit insurance pay insured depositors, see GILLIAN G.H. GARCIA, DEPOSIT INSURANCE: ACTUAL


insured limit.108 Reimbursement of insured depositors does not therefore depend on when or whether the failed bank’s assets are liquidated, or the liquidation value of those assets. Speedy satisfaction of insured depositors is necessary to preserve the liquidity of their deposits. However, speed in liquidating the failed bank’s assets is unnecessary to do so. The FDIC can simply pay the depositors directly or pay a bank to assume the liabilities. As long as depositors have uninterrupted access to their insured deposits, they should be indifferent as to the fate of their depository bank or its assets. Insured depositors therefore have no reason to run on their failed banks. For the same reason, the failure of unhealthy banks cannot justifiably induce insured depositors of healthy banks to run on their banks as long as they believe that the insurance fund is solvent and that they will retain access to their deposits in the event of bank failure.

As insurer of the deposits, the FDIC must be subrogated to their rights so that it can be reimbursed for the expense of making depositors whole. However, the liquidation of the assets need not occur concurrently with the payment of depositors. In fact, the FDIC often retains a sizable portion of the failed bank’s assets, by choice or necessity, and liquidates them over time.109 The FDIC sometimes does transfer assets to the acquiring bank as partial compensation for the acquirer’s assumption of insured deposits and other liabilities. However, as noted above, the acquirer usually receives less than a quarter of the failed bank’s assets,110 and the FDIC uses cash payments to make up the difference between the deposits assumed and the assets received. In other words, the FDIC must have a claim against the assets of a failed bank, but it need not control the disposition of these assets.

FDIC control of assets might be justified if it reduced the size of the FDIC’s claim against the failed bank. For instance, consumers value liquidity and the convenience provided by a bank account. Accordingly, the FDIC can usually find an acquirer willing to assume the insured deposits for less than the amount that it would cost the FDIC to repay the insured depositors in full.111 The FDIC’s control of the physical assets of the failed bank would be appropriate if consumer depositors were more likely to continue their banking relationship with the acquiring bank if the acquiring bank purchases some of these assets. In this case, acquirers of consumer

AND GOOD PRACTICES 83–84 (2000). 108. 12 U.S.C. § 1821(a)(5) (2006). 109. DIV. RESOLUTION & RECEIVERSHIPS, supra note 13, at 22. 110. Supra note 81 and accompanying text. 111. The acquirer typically "pays" a premium for the deposits.

1010 67 WASH. & LEE L. REV. 985 (2010)

deposits would typically prefer to purchase the failed bank’s physical assets.112 But the factual assumption is wrong: Acquirers do not typically pick them up. Instead, the standard purchase and assumption agreement merely gives the acquirer an option to purchase them.113 This is not hard to explain. Because acquiring banks have their own branches and other deposit facilities, they (and their depositors) do not usually need the failed bank’s physical assets. Acquirers can preserve a consumer depositor’s preexisting banking relationship with her failed bank, if at all, without buying the failed bank’s furniture or equipment. In fact, throughout the 1980s the assets sold as a part of the basic purchase and assumption agreement were limited to cash and cash equivalents.114 Acquirers are unlikely to pay more for deposits merely because the FDIC offers the failed bank’s physical assets as well. In any case, physical assets account for a tiny fraction (an average of 2.29%) of the assets of the failed banks.

It might be thought that uninsured depositors and subordinate creditors prefer a quick liquidation of a failed bank’s assets. Their willingness to invest in banks would be jeopardized by delays in the recoupment of their investments. If the FDIC’s control of the resolution process speeds resolution, it could increase the value of their claims.115 This argument is unconvincing. First, most uninsured depositors and other creditors withdraw their deposits and are paid in full before a bank fails and thus do not suffer the consequences of delay.116 The behavior of foreign depositors in

112. A similar argument with regard to the bank’s employees is weak. The acquirer may wish to hire the employees of the failed bank to increase the chance that the bank’s depositors or borrowers will continue their relationship with the bank. However, the FDIC need not control the resolution process to allow this to happen. These employees will undoubtedly have at-will contracts with the failed bank, and the acquiring bank is free to hire them (or not). Moreover, coordinating the hiring of these employees is likely to be relatively easy, as banks have few employees considering their asset size. Nearly 80% (86 of 109) of the banks in our sample had fewer than one hundred employees. 113. See Purchase and Assumption Agreement Between FDIC and Wilshire State Bank, 17 (June 26, 2009), available at http://www.fdic.gov/bank/individual/failed/mirae_ P_and_A.pdf (last visited Sept. 28, 2010) (providing a standard purchase option) (on file with the Washington and Lee Law Review). 114. See Eric Bloecher & John F. Bovenzi, Evolution of the FDIC’s Resolution Practices, in MANAGING THE CRISIS: THE FDIC AND RTC EXPERIENCE 65, 67 (1998) ("[B]ecause only ‘clean’ assets, such as cash and cash equivalents, were passed, due diligence was not required by bidders."). 115. This assumes that the amount received in the sale remains the same. If the market for the failed bank’s assets are not liquid, a quick sale could reduce the liquidation value. 116. See Christine M. Bradley & Lynn Shibut, The Liability Structure of FDIC-Insured Institutions: Changes and Implications, 18:2 FDIC BANKING REV. 1, 20 (2006) ("[M]ost uninsured depositors do not lose money when a bank fails because they manage to withdraw


Continental Illinois Bank and Trust’s failure illustrates the behavior. Deposits in foreign offices are not "deposits" under the FDI Act, and they are therefore uninsured and junior in priority to deposits in domestic offices.117 Foreign depositors exited the bank in droves before it failed.118 Most uninsured depositors therefore are unconcerned about the liquidation value of their claims against a failed bank because they will not have claims when the bank fails.

Uninsured depositors and other creditors who remain and are concerned about a bank’s continued solvency can and do take security interests in the bank’s assets.119 Applicable law sometimes requires certain depositors and lenders to do so in any case.120 Collateral makes uninsured secured creditors generally indifferent to the speed at which the failed bank is resolved because they are almost certain to recover in full. To the extent that they are concerned, there is no reason why they must wait until all of the assets are sold. Uninsured depositors could sell their claims to those who are more patient; by definition uninsured deposits are for substantial amounts. Alternatively, the resolution process could allow partial compensation of the their deposits and receive full payment beforehand."); James A. Marino & Rosalind L. Bennett, The Consequences of National Depositor Preference, 12:2 FDIC BANKING REV. 19, 28–30 (1999) (noting that unprotected depositor and creditor claims are typically small at the time of failure); John S. Jordan, Depositor Discipline at Failing Banks, NEW ENGLAND ECON. REV., MAR.–APR. 2000 at 15 (noting that 70% of uninsured deposits in a sample of failed banks were withdrawn within two years of failure); see also A.M. Davenport & K.M. McDill, The Depositor Behind the Discipline: A Micro-Level Case Study of Hamilton Bank, 30 J. FIN. SERVICES RES., Aug. 2006, at 93 (providing evidence that uninsured depositors respond to the financial health of banks). 117. See 12 U.S.C § 1813(m) (2006) (defining an insured deposit). 118. See Kate McDermott, Continental Illinois National Bank and Trust Company, in MANAGING THE CRISIS: THE FDIC AND RTC EXPERIENCE 545, 547 (1998) ("Large foreign depositors became nervous after hearing rumors of Continental’s imminent failure, and, in May 1984, began a high-speed electronic deposit run on the bank."). 119. See James B. Thomson, The National Depositor Preference Law, ECON. COMMENT., FED. RES. BANK CLEVELAND, Feb. 15, 1994 (examining depositor preference); Eric Hirschhorn & David Zervos, Policies to Change the Priority of Claimants: The Case of Depositor Preference Laws, 4 J. FIN. SERVICES RES. 111, 111–15 (1990) (discussing how depositor preference rules impact the creditors). 120. FHLB advances require collateralization, as do deposits of public funds in most jurisdictions. See, e.g., COLO. REV. STAT. § 11-10.5-107(5) (2009) ("As an ongoing requirement of designation as an eligible public depository, any such depository shall pledge collateral having a market value in excess of one hundred two percent of the aggregate uninsured public deposits."); OR. REV. STAT. § 295.008(2)(b) (2009) (noting conditions for acting as a custodian or bank depository); VA. CODE ANN. § 2.2-4402 (2009) ("Every qualified public depository shall deposit with the State Treasurer, . . . eligible collateral equal to or in excess of the required collateral of such depository to be held subject to the order of the Treasury Board.").

1012 67 WASH. & LEE L. REV. 985 (2010)

uninsured depositors in advance of the completion of the process. In fact, in the past the FDIC has made advance payments to uninsured depositors,121 although it has since discontinued the practice.122 There is even less reason to rush on behalf of the nondepositors. As discussed more thoroughly in Section III.B.2 below, failed banks almost always have insufficient assets to repay the FDIC and uninsured depositors in full.123 Because they are junior to depositors in priority, nondepositors receive nothing. They therefore should not care about the speed at which a failed bank is resolved.

Perhaps the greatest weakness with the speed of liquidation argument is that the FDIC does not, in fact, quickly liquidate the assets of the failed bank. The FDIC’s own resolution manual proposes a four-year liquidation schedule,124 and the average time elapsed between the seizure of failed banks between 2002 and 2003 and the date of the last distribution to depositors was forty-seven months. In only one transaction was the final payment made in less than one year (ten months).

B. The FDIC as Residual Claimant

In some bankruptcies, such as most single-asset real estate cases, one creditor is owed much more than any other. While bankruptcy does apply some special rules in the single-asset real estate context,125 it does not assign total control over the process to the largest creditor. It does not do so because of the fear that the largest creditor will sell the real estate too quickly and for too little at the expense of junior claimants.126 121. See S. Blair Bean, Treatment of Uninsured Depositors and Other Receivership Creditors, in MANAGING THE CRISIS: THE FDIC AND RTC EXPERIENCE 245, 250 (1998) ("If the FDIC’s actual collections on the assets of the failed institutions exceeded the advance payments and administrative expenses of the receivership, the uninsured depositors and other creditors received additional payments on their claims."). 122. See George G. Kaufman, Depositor Liquidity and Loss Sharing in Bank Failure Resolutions, 22 CONTEMP. ECON. POL’Y 237, 245 (2004) (noting the FDIC’s prompt payment of insured deposits when conducting a direct payoff transaction). 123. See infra notes 125–28 and accompanying text (discussing the role of the FDIC as a residual claimant in bank failures). 124. See supra note 13 and accompanying text (providing a timeline for the disposition of a failed bank’s assets). 125. See, e.g., 11 U.S.C. § 362(d)(3) (2006) (governing the stay of an act against single asset real estate). 126. Bankruptcy law also grants minority creditors protection against dominant creditors who share the same priority. Within each class, approval of a plan of reorganization requires the vote of two-thirds of the value of claims and the majority of claims. Id. § 1126(c). As a result, a creditor that is owed 99% of the debt of the debtor may


The law can, however, entrust the largest creditor with control over the process if it is also the residual claimant: A claimant that receives all of the gains associated with a good decision and all of the losses associated with a bad decision. A creditor is the residual claimant if two conditions are satisfied: i) those who have equal priority with the creditor have negligible claims or the creditor is fully compensated for the costs it incurs in maximizing the value of assets, and ii) regardless of the decisions made by the creditor, those senior to the creditor in priority will be paid in full and those junior to the creditor will not be paid anything. In most nonbank insolvencies there is no claimant who fully meets each criteria and the law must allocate control among the parties who bear at least some of the risk associated with the decision.

Consider what happens when the decision-maker is not the residual claimant. Take the first condition. Assume that a firm has two creditors of equal priority and each are owed $50. Assume that the firm’s assets can be sold for $80 if the decision-maker incurs a cost of $15. If this cost is not incurred, the firm can only be sold for $60. It is socially efficient for the decision-maker to incur this cost, because the sale nets $5 more if the decision maker incurs the cost ($80-$15=$65) than if she does not incur it ($60–0=$60). However, she will not incur the cost unless she is reimbursed. If the decision maker does not incur the cost, she will receive $30 (half of $60). If she incurs it, she will receive a net of $25 ($40-$15). This problem can be eliminated by allowing the decision-maker to recoup her administrative costs before dividing the proceeds with the other creditor, but this can create its own problems. If the decision-maker derives some private benefit from these expenses (perhaps she can overstate her expenses), she can use this priority to divert value from the other creditor. However, her ability to do so declines as the proportion of the debt owed to the decision-maker increases.

Now consider the second assumption. Assume that a firm again has two creditors that are owed $50, but now also assume that one is senior to the other. Suppose the decision-maker has three options. If she adopts plan A, there is a 90% chance that the firm’s assets will be sold for a present value of $70 and a 10% chance that they will be sold for a present value of $20. If she adopts Plan B, there is a 50% chance that the firm’s assets will be sold for a present value of $90 and a 50% chance that they will be sold for a present value of $20. Plan C would be to sell the firm’s assets for $60 with certainty.

be unable to approve a plan of reorganization if there were multiple smaller creditors and they opposed the plan.

1014 67 WASH. & LEE L. REV. 985 (2010)

Plan A is clearly the socially optimal choice as its expected present value of $65 exceeds that of Plan B ($55) and Plan C ($60). However, if the decision-maker holds the junior claim, she will prefer Plan B as this will provide her with an expected payment of $20 which is more than what she will expect to receive from Plan A ($18) or Plan C ($10). If the decision-maker holds the senior claim, she still does not have the right incentives. The senior claimant will prefer Plan C as this provides her with an expected payment of $50 which is more than what she would receive from Plan A ($47) or Plan B ($35).127 There is no true residual claimant in this hypothetical. Note that this problem is not inevitable in the presence of senior and junior claimants; it disappears if we change the value of the claims. Assume, for example, that the senior claim is owed just $20 and the junior claim is owed $80. The senior claim will be paid in full regardless of the plan chosen. The junior claim bears all of the risk (it is the residual claimant) and would choose the plan (Plan A) that maximizes social welfare. Similarly, if we assume that the senior claimant is owed $90 and that the junior claimant is owed $10, then the junior claimant will receive nothing regardless of the plan adopted. The senior claimant bears all of the risk and would make the choice that maximizes social welfare.

Allocating decision-making authority to the residual claimant is generally defensible. However, to supply ex ante efficient incentives, an exception may occasionally justify giving decisions to a junior claimant. The residual claimant has an incentive to make ex post socially optimal choices, because it benefits from doing so. But it might be insufficiently motivated to make decisions at earlier stages that avoid financial risk later on. Additional motivation is provided by taking decisions away from the residual claimant and giving them to junior claimants when the firm is in financial distress. Because junior claimants lack the incentive to make decisions to benefit residual claimants, the shift in decision making potentially harms residual claimants’ interests. For this reason, residual claimants have a further incentive to remain in control of the firm’s fortunes. To remain in control, they must make decisions that avoid putting the firm into financial distress. An ex post inefficient allocation of decision making is required to give residual claimants ex ante efficient incentives.

Assume, for example, that the senior secured creditors are the residual claimants of a failed firm and that allocating decision-making rights to

127. These examples ignore the possibility of Coaseian bargaining. For example, the junior claimant in this scenario could simply repay the senior claimant in full and thereby obtain the right to make the decisions.


unsecured creditors would result in a social loss (and a loss to the senior secured claimant) of $30. While this creates an ex post cost, it may provide ex ante incentives for the senior secured lender to curb excessive risk taking by the debtor. Assume that one year before insolvency the firm could have sold its assets for $100 and that the only debt was $50 owed to the senior secured creditor. Finally, assume that the debtor also had the choice of engaging in a project that, if successful, would have been worth $200. However, the project would fail half of the time, and if it did so the firm’s assets would be worth just $50 and the firm would incur tort claims of $100. Society would not want the firm to undertake this project. The expected value of the project is $75 ((.5 x $200) + (.5 x -$50) = $75), while liquidation of the firm would net $50 for the unsecured creditors ($100-$50). However, the project would make the shareholders better off, on average, than if the firm were liquidated. If the firm were liquidated, shareholders receive nothing. For their part, the senior secured lender would have no reason to stop the shareholders from pursuing the project because its priority guarantees it payment in full ($50) whether the project is undertaken or the firm liquidated. Note that the lender would oppose the project if either it shared control upon insolvency (and thus $30 was wasted) or if it were subordinated to the tort victims. Imposing the ex post cost is a way to avoid the inefficiencies caused by the tort claimant’s lack of priority and ability to control the debtor’s investment decisions. The exception that separates decision making and residual claimants is inapplicable in bank failures. As we show below, there are very few fixed claimants that are junior to the FDIC; general creditors and subordinated creditors do not account for a meaningful component of the capital structure of failed banks.128

Two pieces of evidence suggest that the FDIC is truly the residual claimant in the overwhelming majority of bank insolvencies. First, in Section III.B.1 we show that the FDIC accounts for the vast majority of claims on the assets of both healthy and failed banks. The FDIC’s status as the holder of the overwhelming majority of the debt does not ensure that it will be the residual claimant. Whether the majority creditor is a residual claimant depends on the value of the assets available for distribution to creditors. However, the greater the share of debt held by a creditor and the greater the value of the debt relative to the equity, the greater is the chance that the majority creditor will be the residual claimant too. For example, consider a bank with $10 in senior secured

128. Infra Table 1.

1016 67 WASH. & LEE L. REV. 985 (2010)

claims, $80 in insured deposits, and $10 in general unsecured claims. Because the secured claims will have priority over the deposits and the deposits will have priority over the general claims, the FDIC will be the residual claimant as long as the plausible range of asset values is between $10 and $90. If assets have a $90 value, the FDIC (subrogated to the rights of insured depositors) is repaid in full ($90-$10=$80). The FDIC receives nothing if assets have a value of $10 or less ($10-$10=$0). The FDIC is repaid in part if asset values are between $11 ($11-$10=$1) and $89 ($89-$10= $79).

Second, we examine the actual payouts from failed banks in Section III.B.2. We show three patterns in these payouts: i) the most important secured creditors have recovered in full in every bank failure to date; ii) the FDIC suffered significant losses in the vast majority of bank failures; and iii) general creditors almost never receive any dividends. As a result, it is the FDIC that would enjoy the gains from a distribution method that yields greater proceeds.

1. The FDIC as the Largest Creditor

Most banks are part of holding company structures. However, with the possible exception of the very largest banks, discussed below, the capital structures of the actual banks are extremely simple. Domestic deposits account for the overwhelming majority of bank liabilities, and almost all of these domestic deposits are insured. These banks have comparatively little general unsecured debt and almost no subordinated debt. This is true whether we examine failed banks or banks more generally.


Table 1: Liability Structure of Failed Banks: 1/1/1995–5/31/2009 Less than

$100 million

$100 million to

$500 million

$500 million to $1 billion

$1 billion to $5

billion

More than $5 billion

Banks 44 45 10 15 5 Deposits as Percent of Liabilities (Std. Dev.)

96.25%

(4.61%)

92.85%

(7.33%)

89.13%

(5.38%)

87.35%

(10.16%)

70.39%

(7.73%) Percent of Deposits Insured (Std. Dev.)

89.81%

(10.47%)

80.75%

(18.13%)

66.03%

(23.37%)

81.01%

(16.93%)

84.35%

(6.15%) Foreign Deposits as Percent of Liabilities (Std. Dev.)

0.00%

(0.00%)

0.35%

(1.62%)

0.00%

(0.00%)

0.35%

(1.35%)

0.00%

(0.00%) FHLB as Percent of Liabilities (Std. Dev.)

1.43%

(3.55%)

3.94%

(5.73%)

8.82%

(5.62%)

7.04%

(9.95%)

26.10%

(7.91%) Repurchase obligations as % of Liabilities (Std. Dev.)

0.35%

(1.12%)

1.21%

(3.26%)

0.49%

(0.93%)

3.47%

(6.60%)

0.12%

(0.17%) Subordinated debt as Percent of Liabilities (Std. Dev.)

0.07%

(0.30%)

0.00%

(0.00%)

0.15%

(0.43%)

0.18%

(0.38%)

0.62%

(1.22%) General Unsecured Claims as Percent of Liabilities (Std. Dev.)

1.73%

(2.59%)

2.01%

(3.48%)

1.41%

(1.32%)

2.20%

(2.36%)

2.77%

(4.00%)

Table 1 documents key characteristics of the liability structure of failed banks as of the last reporting period before their failures. For now we will ignore the very largest banks—those with assets greater than five billion dollars. In all other banks, deposits account for the overwhelming majority of liabilities. The proportion declines somewhat according to the size of bank, but remains high. Deposits constitute 96% of liabilities in failed bank with

1018 67 WASH. & LEE L. REV. 985 (2010)

assets below $100 million.129 The percentages for banks with assets between $100 and $500 million, $500 million and $1 billion, and $1 billion and $5 billion are 93%, 89%, and 87%, respectively. Out of the 120 banks that failed between January 1, 1995 and May 31, 2009, only ten had deposits that constituted less than 80% of liabilities.130 Healthy banks have a similarly high percentage of deposits.131

The FDIC is only subrogated to the insured deposits, but Table 1 reveals that a strong majority of deposits are insured. The percentage of insured deposits at healthy banks is a little lower.132 Table 1 presents data from the last Call Report issued before the bank’s failure, and uninsured deposits are likely to decline further as the bank nears insolvency. If the resolution process is not yet complete, the FDIC reports the claims of uninsured depositors when the acquiring bank assumes only the insured deposits. In the sixteen transactions for which we could find data, uninsured depositors averaged just 4.75% of total deposits. Uninsured depositors typically either exit before a bank fails or convert their uninsured amounts to insured deposits. The bank replaces the exited funds with another funding source. For instance, a study of failed New England banks found that they lost about 70% of their uninsured deposits within two years of failure.133 The banks, however, replaced these deposits with insured deposits. Such replacement allows the banks to retain the same source of funding while increasing the percentage of insured deposits. In addition, failed banks tend to overstate the amount of uninsured deposits, thereby understating insured deposits.134 By 129. Asset values of banks are adjusted for inflation. 130. The predominance of deposit liabilities has not changed much over time. Table 1 would not change materially if we focused only on banks that failed between 2008–2009. 131. See FDIC, Statistics on Depository Institutions Report, http://www2.fdic. gov/sdi/main.asp (last visited Sept. 28, 2010) (revealing that as of Dec. 31, 2008 deposits approximately 83% of liabilities for banks with assets up to $100 million, 90% for banks with assets between $100 million and $1 billion, and 73% for banks with $1 billion or more in assets) (on file with the Washington and Lee Law Review). 132. See id. (revealing that as of Mar. 31, 2009, approximately 78% of deposits are insured at banks with up to $100 million in assets, 71% at banks with between $100 million and $500 million in assets, and 59% at banks with more than $1 billion in assets). 133. See John S. Jordan, Depositor Discipline at Failing Banks, NEW ENGLAND ECON. REV., Mar.–Apr. 2000, at 15, 16 ("[F]ailing banks in New England experienced a 70 percent decline in their uninsured deposits in their final two years of operation."); Matthew T. Billett et al., The Cost of Market Versus Regulatory Discipline in Banking, 48 J. FIN. ECON. 333, 333 (1998) (finding that banks increase their proportion of insured deposits in response to a ratings downgrade). 134. See ANDREW M. DAVENPORT ET AL., AN EVALUATION OF THE DENOMINATOR OF THE RESERVE RATIO: FDIC STAFF STUDY 16 (2007), available at http://www.fdic.gov/ deposit/insurance/initiative/Denominator_Board.pdf (examining a sample of failed banks to


contrast, healthy banks, which do not experience a run on uninsured deposits, retain higher percentages of uninsured deposits.

Uninsured domestic deposits are equal in priority to the insured claims to which the FDIC is subrogated. Other claims are either senior or junior to the FDIC, but these claims are rarely significant. General claims rarely account for more than 5% of a bank’s assets.135 Foreign deposits are junior to the FDIC’s claims,136 but just three banks in our sample had any foreign deposits, and none had foreign deposits that accounted for more than 10% of all deposits. With the possible exception of the very largest banks, discussed below, this is true of healthy banks as well.137 Small banks generally do not have foreign branches.

The comparative absence of subordinated debt among failed banks deserves notice.138 Subordinated debt represents only 1% of total liabilities among every category of bank, and is literally nonexistent in failed banks with less than $500 million in assets. The pattern is consistent with the issuance of subordinated debt by U.S. commercial banks generally. In the 1990s, the percent of banks issuing subordinated debt declined for banks of all sizes except for the very largest banks.139 Subordinated debt was present on the balance sheets of almost all of the largest banks in 1998.140 This trend continues. One study found that only 9% of banks sampled between 1996 and 2005 issued any subordinated debt, and nearly two-thirds of these banks

reveal an understatement in final Call Reports). 135. We estimate general claims as the sum of "Trading Liabilities," "Other Borrowed Liabilities" and "Other Liabilities" less "FHLB Advances." 136. See 12 U.S.C. § 1821(d)(11)(A)(2006) (espousing the priority for liquidation disbursements); id. § 1813(l) (defining the term "deposit"). 137. See supra note 25 and accompanying text (displaying the liability structure of failed banks). 138. As discussed in Part IV, subordinated debt is more common at the holding company level. About 80% of U.S. banks and almost all banks with assets of at least one billion dollars are owned by holding companies. See FEDERAL RESERVE SYSTEM STUDY GROUP ON SUBORDINATED NOTES AND DEBENTURES, STAFF STUDY 172: USING SUBORDINATED DEBT AS AN INSTRUMENT OF MARKET DISCIPLINE at 26 (1999) (displaying the subordinated debt issuance by insured commercial banks from 1991–98). About 10% of these holding companies issue subordinated debt with the issuers concentrated among the very largest holding companies. Id. at 26. The percentage of bank holding companies issuing such debt has declined for every size bank holding company other than large companies (although the amounts have increased). Id. In any case, our story is about the capital structure of the bank, not the bank holding company. The FDIC resolves failed banks through receivership; bank holding companies file for bankruptcy. 139. Id. 140. Id.

1020 67 WASH. & LEE L. REV. 985 (2010)

were large banks with assets exceeding 1 billion dollars.141 Thus, subordinated debt is as relatively rare in healthy banks as it is in failed banks.142 The liability structure of healthy banks is similar in these respects to that of failed banks. Failed banks in all asset size categories show little debt owed to general creditors and almost no subordinated debt.

Secured claims are senior to the FDIC, and some domestic deposits are secured by the bank’s securities or mortgages.143 Federal Home Loan Bank advances are the most important type of secured liability for many banks.144 These advances must be at least fully collateralized by securities.145 Although many failed banks have not taken such advances, Federal Home Loan Bank advances are significant liabilities for banks which have taken them. Table 1 shows that Federal Home Loan liabilities rise with bank size from about 1.4% for the smallest banks to almost 10% for banks between one and five billion and over 26% for banks over five billion. We note that the importance of FHLB loans may be due in part to the real estate boom of the first decade of the twenty-first century. In sum, Table 1 shows that failed banks other than mega-banks have a comparatively simple liability structure. As banks near failure, most of their liabilities are domestic deposits, and the overwhelming majority of failed banks have no significant foreign deposits, general unsecured claims or subordinated debt. A few banks do have some

141. See A. Sinan Cebenoyan & Fatma Cebenoyan, Subordinated Debt, Uninsured Deposits, and Market Discipline: Evidence from U.S. Bank Holding Companies 1, 17 (Hunter College Dept. of Econ. Working Papers, Paper No. 421, July 2007) (providing statistical information regarding the issuance of subordinated debt). 142. See Douglas D. Evanoff & Larry D. Wall, Subordinated Debt and Bank Capital Reform 13, 39 (Fed. Res. Bank of Chicago Working Papers, Paper No. 2000-24, 2000) (noting the frequency with which banks issue subordinated debt). 143. State statutes often require deposits by state and political subdivisions to be collateralized. A random sample of failed banks by asset size reveals that these "preferred deposits" range between 0% and 3% of a bank’s deposit liabilities. Such collateralization therefore represents a relatively unimportant sort of security interest. 144. See Rosalind L. Bennett et al., Should the FDIC Worry About the FHLB? The Impact of Federal Home Loan Bank Advances on the Bank Insurance Fund 29 (Fed. Res. Bank of Richmond, Working Paper No. 05-05, 2005) (stating that FHLB advances account for close to half of secured funding at U.S. banks as of 2003, amounting to almost 3% of total bank assets). 145. See 12 U.S.C. § 1430(a)(3) (2006) ("A Bank, at the time of origination or renewal of a loan or advance, shall obtain and maintain a security interest in collateral . . . ."). Federal funds and repurchase agreements are also an important type of secured liabilities. See Lynn Shibut, Should Bank Liability Structure Influence Deposit Insurance Pricing? 14 (FDIC, Working Paper No. 2002-01, 2002) ("The most prevalent types of secured credits include FHLB advances, repurchase agreements, public deposits and borrowings from Federal Reserve Bank (FRB) discount window.").


Federal Home Loan Bank advances or other secured debt, but these secured claims rarely account for more than 25% of all liabilities. Some banks do have significant amounts of uninsured deposits (at least at the time of the last call report before failure), and these depositors will have a claim on the assets separate from the FDIC. However, the FDIC will share proceeds on a pari-passu basis with these creditors, and so they can generally trust the FDIC to maximize recovery as long as the FDIC is appropriately compensated for its expenses and is trying to minimize the loss to the insurance fund.

2. FDIC Losses on Failure

Section III.B.1 demonstrates that insured deposits account for the overwhelming majority of the liabilities of insolvent banks. It also demonstrates that most of the remaining liabilities will be senior to the FDIC’s claim on the failed firm’s assets (FHLB loans) or will share in the proceeds on a pari-passu basis (uninsured deposits). The senior secured liabilities should not disturb the FDIC’s status as residual claimant because they are always paid in full; there has never been a default on an FHLB loan.146 The uninsured deposits should not disturb the FDIC’s status as residual claimant because most uninsured deposits will flee as the bank nears insolvency and the remaining will share in the proceeds on a pari-passu basis. The presence of a large amount of claims junior to the FDIC could distort the FDIC’s incentives, but these claims (foreign deposits, general unsecured claims and subordinated debt) are practically nonexistent in nearly all bank failures. Moreover, these claims rarely receive any payment. Of the fifty-one receiverships begun between 1995 and 2008 for which the FDIC reports the percentage distribution to general claimants, general claimants have received nothing in forty-nine.147

146. See Mark J. Flannery & W. Scott Frame, The Federal Home Loan Bank System: The "Other" Housing GSE, 91 FED. RES. BANK OF ATLANTA, ECON. REV. 33, 39 (2006) ("No FHLB has ever suffered a credit loss on an advance."). 147. General claimants received 28% of their claims in the failure of Net First National Bank and 100% of their claims in the failure of Dollar Savings Bank. The distribution to general creditors is only available for receiverships that are incomplete. It is therefore possible that general creditors could receive distributions in more of these receiverships.

1022 67 WASH. & LEE L. REV. 985 (2010)

Even if the FDIC were the sole creditor of the bank, it still would not be the residual claimant if the value of the assets exceeded the value of the bank’s liabilities. In theory, a bank with a solvent balance sheet could fail if its regulator mistakenly declared the bank insolvent and seized it or if a liquidity shock rendered the bank unable to meet short term obligations and unable to issue new shares to raise new funds. In practice, however, a failed bank’s assets almost never exceed its liabilities: the FDIC almost always loses significant amounts of money. As long as the FDIC is losing money, it has the proper incentive to maximize the amount recovered, as the FDIC will receive each additional dollar raised.148 Figure 2 shows the frequency with

148. Strictly, the FDIC has an incentive to maximize amounts recovered when its distribution of amounts recovered exceeds the cost to it of the recovery. The FDIC’s distribution is limited because the amounts recovered must be shared with uninsured domestic depositors. See 12 U.S.C. § 1821(d)(11)(A)(ii) (noting depositor ordering preference). Because it incurs costs in recovering assets, the FDIC’s incentive therefore is to maximize net returns to it, not total returns. The qualification usually is unimportant. There are relatively few uninsured depositors of failed banks. See supra Table 1 (displaying the liability structure of failed banks). Applicable rules give priority to the expenses the FDIC incurs in disposing of assets over depositor claims. As a result, it retains the incentive to maximize the value of the assets.

The FDIC may not want to sell the failed bank’s assets for their greatest value if doing so would jeopardize the financial stability of the acquiring bank and the acquirer is also FDIC- insured. Instead, given some assumptions about liquidity, the FDIC may wish to use the sale to subsidize the acquiring bank. We are unconcerned about this theoretical

0.00%

10.00%

20.00%

30.00%

40.00%

50.00%

60.00%

0 to 5% 5 to 10% 10 to 20 More than 20%

Figure 2: Estimated FDIC Losses as Percent of Deposits (1/1/1995-5/31/2009)

Percent of Transactions


which FDIC losses (expressed as a percentage of deposits)149 fall into various categories. Note that the FDIC’s loss rate is typically very high—more than half (56%) of bank failures resulted in FDIC losses exceeding twenty-percent of total deposits. The loss rate is less than 5% in just 14% of transactions.

The FDIC’s status as a residual claimant justifies its control over the resolution process. We do not claim that Congress gave the FDIC control over the resolution process because the FDIC is the residual claimant on most failed banks’ assets. This historical claim is false. Instead, the FDIC’s status as a residual claimant in control of bank resolutions likely was an unintended consequence of Congress’s grant of liquidation powers to the FDIC. Congress created a federal receiver with powers to liquidate failed banks and gave the FDIC these powers. Federal deposit insurance, along with the right of subrogation to insured depositors, in turn made the FDIC the residual claimant in most banks being liquidated. But Congress did not give the FDIC liquidation powers because it was the residual claimant. The drafting history of the legislation creating the FDIC and federal deposit insurance supports this conclusion.

The Banking Act of 1933 established both the FDIC and deposit insurance. The year before the Act was enacted, Senator Carter Glass, one of its architects, introduced in the Senate Banking and Currency Committee a proposed banking bill. The bill created a "Federal Liquidating possibility for three reasons. First, the possibility is not realistic. The FDIC requires bidders of failed banks to meet capital and managerial requirements. An acquirer whose acquisition would put it in financial jeopardy is not a qualified bidder. Second, under the least-cost resolution standard binding on the FDIC, the FDIC arguably lacks the legal authority to make the sale proposed. Section 1823(a)(4)(ii) prohibits the FDIC from making expenditures or incurring obligations with respect to the failed bank unless doing so has the least cost to the deposit insurance fund among available alternatives. 12 U.S.C. § 1823(a)(4)(ii). Selling assets of an insolvent bank to an acquirer at below value increases the loss to the failed bank’s creditors, including the bank’s insured depositors. Because the FDIC must pay insured deposits from the deposit insurance fund, the increased loss from a below-value sale is borne by the deposit insurance fund. The alternative of selling assets at value makes a sale at value a less costly alternative than a below-value sale. The least-cost resolution standard therefore bars below-value sales. Third, if the FDIC is truly the residual claimant of the failed bank (as the data we present suggest it almost always is), the failure to maximize asset values comes at the expense of the deposit insurance fund. It would not harm other creditors of the failed bank. Thus, even if the FDIC has the legal authority to subsidize the acquiring bank, it will sell assets below value only if doing so is socially efficient. 149. The literature often presents FDIC losses as a percentage of bank assets. We choose instead to present them as a percentage of deposits because the FDIC does not insure the bank’s assets directly. Thus, losses divided by deposits (or insured deposits) give a more accurate sense of the rate of loss. We use deposits instead of insured deposits because it is hard to estimate insured deposits reliably.

1024 67 WASH. & LEE L. REV. 985 (2010)

Corporation."150 Glass’s bill gave the Corporation the power to liquidate failed banks by purchasing their assets for the purpose of speeding up payments to depositors.151 The corporation in turn would sell the assets and remit the sale proceeds to the bank’s receiver. The receiver in turn would distribute them to depositors. As proposed, the liquidation corporation would control the failed bank’s assets; depositors would be repaid from the sum the corporation received for the assets. Glass’s bill did not propose to insure deposits. Depositors therefore would retain the risk of the bank’s failure. They suffered loss for amounts above their proportionate share of the failed bank’s assets. The Federal Liquidating Corporation’s role simply was to quicken receipt of available distributions to depositors. Because the corporation did not insure deposits, it had no claim (derivative or original) on the failed bank’s assets.

Carter’s bill failed over objections principally to the capitalization of the proposed liquidating corporation.152 However, it became the template the Roosevelt administration and Congress used in drafting the Banking Act of 1933.153 This meant that the Banking Act in its initial draft form gave the power to liquidate failed national banks to a single resolution authority. Prior law gave the Comptroller of the Currency the authority to appoint a receiver for failed national banks. By 1932 most states had given their bank

150. See U.S. Senate, Banking and Currency Committee, 71st Cong., Section 10 (1932), reproduced in Federal Reserve Board’s Comments and Recommendations on the Glass Bill (S. 4115), 18 Fed. Res. Bull. 206, 211 (1932). See generally Draft of S.3315, Section 9, Carter Glass Papers, Alderman Library, Special Collections, University of Virginia, Box 284. 151. See RIXEY SMITH & NORMAN BEASLEY, CARTER GLASS: A BIOGRAPHY 305–06 (1st ed. 1939) (describing the bill proposed by Glass); H. PARKER WILLIS & JOHN M. CHAPMAN, THE BANKING SITUATION: AMERICAN POST-WAR PROBLEMS AND DEVELOPMENTS 65–66 (1st ed. 1934) (same). 152. See Federal Reserve Board’s Comments and Recommendations on the Glass Bill (S. 4115), 18 Fed. Res. Bull. 206, 211–25 (1932) (containing evaluations of Federal Reserve Board and Federal Advisory Council). 153. See SMITH & BEASLEY, supra note 151, at 65–66 (describing the adoption of the Banking Act of 1933). Roosevelt found Carter’s proposed Federal Liquidating Corporation attractive. Although Roosevelt initially wanted the Corporation to resolve failed banks by recapitalizing them, he apparently agreed that the Corporation’s role in resolution should be limited to liquidation. H. Parker Willis, who helped Glass produce the 1932 bill, reported to Glass his meeting with Roosevelt: "I explained to him the liquidating provision of the Glass bill and contrasted it with those of the bill proposed by the Comptroller of the Currency and he said that he preferred greatly the Glass provisions because they eliminated the costly, long-drawn receiverships, to which he said he had always felt a strong opposition." Letter of H. Parker Willis to Carter Glass, Nov. 19, 1932, Carter Glass Papers, Alderman Library, Special Collections, University of Virginia, Box 274.


superintendents the same powers.154 The 1933 Act created a resolution authority with the power to liquidate assets of member banks of the Federal Reserve. This gave it powers lacking in a few states and enabled the accumulation of expertise in liquidation through more concentrated control of the resolution process. More significant was the status of the receiver in the Act’s initial draft form.155 Because Glass’s bill did not provide for deposit insurance, the receivership authority under the Act also would have no interest in the assets of banks it was liquidating. This changed in the bill Glass reintroduced in Congress.156 As part of the compromise needed to pass the Banking Act, the Act created federal deposit insurance, and the FDIC established federal deposit insurance. It also gave to the FDIC the power already described in the Act to a liquidating corporation to dispose of assets of any failed FDIC-member bank.157 The Banking Act required the FDIC to pay off insured depositors by creating a temporary national bank, to which insured deposits were transferred.158 Subrogated to the rights of insured depositors, the FDIC now had an interest in the assets it was liquidating. In the draft bill that developed into the Banking Act, the receivership authority was defined first. Only later was the FDIC created and given that authority. There is no evidence that the FDIC was given the authority because it had an interest in the assets it would be liquidating.

IV. Limits of the Case for FDIC Control

Section III argues that the FDIC is likely the residual claimant in nearly all bank failures. It concludes that the FDIC therefore should have the control of the resolution process bank insolvency law currently gives it. Section II’s case for FDIC control depends on at least four assumptions that might not always hold. One assumption is that the capital structure of the failed banks is not a direct result of the fact that the FDIC controls the

154. See UPHAM & LAMKE, supra note 103, at 35–38 ("In 1933 nearly all of the states had adopted the federal rule and had given the state bank supervisor control over closed bank liquidations."). 155. 77 Cong. Rec. 196 (1933); see S.245, 73d Cong. § 7 (1933) (discussing the proposed Act). 156. See 77 Cong. Rec. 3109 (1933) (reintroducing the bank bill through major modifications to the bank bill introduced previously). 157. See Banking Act of 1933, Pub L. No. 73-66, § 8(1), 48 Stat. 162, 172–74 ("It shall be the duty of the Corporation to realize upon the assets of such closed bank . . . ."). 158. See id. at 173–74 (outlining the process for paying off insured depositors of a failed bank).

1026 67 WASH. & LEE L. REV. 985 (2010)

resolution of failed banks. However, the resolution process should, in theory, affect the capital structure firms adopt.159 If, for example, investors expect the FDIC to exercise its resolution powers in a way that harms subordinated debt or general claims (perhaps selling the bank’s assets too quickly and too cheaply), investors will be reluctant to purchase this debt and the firm will rely more heavily on deposits or equity financing. Because a firm’s capital structure is endogenous, we cannot claim that the current resolution system is truly optimal because we cannot directly assess the alternatives. We claim only that our current resolution system appears well-suited to the capital structure that firms have in fact chosen.

A second assumption is the absence of agency costs within the FDIC. If the interests of the FDIC and its employees diverge systematically, the FDIC’s control of the resolution process does not guarantee resolutions that maximize asset values. Agency costs plague all large organizations, but public organizations like the FDIC may be less able to incentivize their agents. While the FDIC control will lead to some sub-optimal resolution decisions, we suggest in subsection A that FDIC control may be the best available alternative. The third assumption concerns liability structure. The liability structure of the very largest banks looks very different from that of the vast majority of failed banks, and the FDIC might not be the residual claimant should these very large banks fail. In this case Congress could subject these banks to special resolution rules that weaken the FDIC’s control. Subsection B does not recommend such rules because of concern that banks may manipulate their capital structure to trigger these rules. A fourth assumption is that a resolution process is needed and that someone must control this process. A number of commentators have suggested changes in the capital structure of banks, primarily the increased use of convertible debt that could make bank failure far less common. We do not address these proposals in detail because we take the existing capital structure of banks as given. Even if the convertible debt is made part of a bank’s capital structure, an insolvency regime still is needed where the value of convertible securities is insufficient to return the failed bank to financial health. Others propose conversion mechanisms that could, in theory, resolve failed institutions "automatically," without significant involvement by the FDIC or a court. Subsection C argues even an automatic conversion

159. This is a standard application of the Modigliani-Miller Irrelevance Theorem. See Franco Modigliani & Merton H. Miller, The Cost of Capital, Corporation Finance and the Theory of Investment, 48 AM. ECON. REV. 261, 278 (1958) (showing that optimal capital structure can depend on bankruptcy costs).


mechanism would still usually leave the FDIC with a substantial role in the resolution of failed banks.

A. Agency Costs in Bank Resolutions

Section III claims that the FDIC is the residual claimant and thus would maximize social value if it followed the law’s command that it exercise its authority in a way that results in the least cost to the deposit insurance fund.160 However, the FDIC does not (and cannot) actually exercise any authority. Its employees must act on its behalf, and these employees will have interests that diverge from those of the primary stakeholders in the FDIC—the member banks and the taxpayers. The employees’ interests likely include a desire for income, leisure and career advancement, in some combination.161 Because FDIC employees’ salaries are fixed, the outcome of a bank resolution has no effect on their income. It might also not affect any of their other interests. Accordingly, FDIC employees might act to serve their interests even at the expense of the FDIC’s interest in resolving a failed bank. The divergence of the FDIC and its employees’ interests creates agency costs.162 Agency costs may prevent the FDIC’s employees from making decisions that minimize the cost to the insurance fund and maximize social welfare. They might, for example, engage in too little marketing efforts and thereby sell the assets for less than they are worth. Alternatively, they might incur excessive costs in managing the disposition of assets.

Intra-organizational agency costs are not unique to the FDIC; large private organizations face these same costs. We might, however, expect private organizations to be better able to control these agency costs. The FDIC may be less able to design its employment contracts to align its

160. See 12 U.S.C. § 1823 (c)(4)(A) (2006) (requiring the FDIC to implement the lowest costing resolution). "Public interest" here is understood broadly. The deposit insurance fund is funded by assessment against participating banks, not by tax revenues. However, bank resolutions that minimize costs to the deposit insurance fund make it unlikely that tax revenues will be used to subsidize failed banks. More generally, resolving failed banks maintains the integrity of the banking system. Both matters are concerns of "public interest." 161. Cf. Arnoud W.A. Boot & Anjan V. Thakor, Self-Interested Bank Regulation, 83 AM. ECON. REV. 206, 208 (1993) (assuming that bank regulators maximize a combination of good supervisory reputation and social welfare). 162. Agency costs are the sum of monitoring and bonding costs, and residual loss to the FDIC when its employees act on its behalf. See generally Michael C. Jensen & William H. Meckling, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, 3 J. FIN. ECON. 305 (1976).

1028 67 WASH. & LEE L. REV. 985 (2010)

employees’ interests with its own,163 and the FDIC may be subject to political pressure164 or lobbying efforts on behalf of special interest groups. A number of scholars who have examined banking regulation prior to insolvency suggest that agency costs have resulted in lax supervision of the soundness of banks or thrifts.165 As a remedy, they propose reforms that would shift more risk to the private sector and thereby induce private sector actors to monitor the debtor or provide more accurate measures of risk.166 These arguments presume that the private organizations that assume this risk would be better able to solve the agency problems and therefore better able to monitor the debtor. If private sector actors are better able to monitor a bank’s activity before insolvency, they may be better able to dispose of a bank’s assets after it becomes insolvent. If public regulators are better monitors than dispersed creditors holding limited debt, they may have an advantage at disposing of these assets. We do not engage the debate over the relative merits of public versus private decision-making.167

The presence of agency costs does not doom the argument for FDIC control. This is because agency costs are inevitable with the sort of decisions

163. See Christopher James, The Losses Realized in Bank Failures, 46 J. FIN. 1223, 1224 (1991) (explaining that "the FDIC may have less of an incentive to maximize the value of assets in liquidation than private market participants since the compensation of government liquidators does not depend on the outcome of their actions"). 164. See, e.g., Edward J. Kane, Principal-Agent Problems in S&L Salvage, 45 J. FIN. 755, 755–56 (1990) (explaining that the FDIC is a potential scapegoat "that politicians may conveniently plan to blame for any scandals or apparent cost overruns that emerge"). 165. See Edward J. Kane, The Incentive Incompatibility of Government-Sponsored Deposit-Insurance Funds, in THE REFORM OF FEDERAL DEPOSIT INSURANCE: DISCIPLINING THE GOVERNMENT AND PROTECTING TAXPAYERS 144, 147 (James R. Barth & R. Dan Brumbaugh, Jr. eds., 1992) ("The limited life observed for government-sponsored funds is less a matter of bad economic luck or specific ‘mistakes’ in regulatory management than of generic principal-agent problems that support structural imbalances in their information, monitoring, enforcement, and incentive systems." (citations omitted)); KANE, supra note 104, at 1 (explaining that "congressional procedures for budgeting and for overseeing the operations of the deposit-insurance bureaucracy made the regulatory strategy of cover-up and deferral practically irresistible"); Asli Demirguc-Kunt, Deposit-Institution Failures: A Review of the Empirical Literature, 25:4 ECON. REV. 2, 2 (1989), available at http://www.clevelandfed. org/research/review/1989/89-q4-DemirgucKunt.pdf ("This dramatic increase in the bank failure rate has intensified public criticism of deposit-institution regulators, since bank safety and soundness is a major regulatory responsibility." (footnote omitted)). 166. See generally FEDERAL RESERVE BANK OF CHICAGO, MARKET DISCIPLINE ACROSS COUNTRIES AND INDUSTRIES (Claudio Borio et al. eds., 2004); Mark J. Flannery, The Faces of "Market Discipline", 20 J. FIN. SERVICES RES. 107 (2001). 167. See MATHIAS DEWATRIPONT & JEAN TIROLE, THE PRUDENTIAL REGULATION OF BANKS 29–45, 193–216 (Mass. Inst. Tech. 2d prtg. 1999) (1993) (discussing the differences between public and private decision-making and between public and private regulation).


that must be made by the FDIC. Decisions about the timing and terms of asset dispositions are complex. They require discretion, whether the decision maker is the FDIC or another authority. Oversight of these decisions can eliminate some self-interested choices by the FDIC’s employees, but perhaps at the expense of the interests of the member banks and the taxpayers. After all, the member banks and the taxpayers cannot effectively dispose of the assets themselves. For this reason, over a range of decisions, there is a tradeoff between the authority to decide how to dispose of assets and agency costs. At the margin the optimal amount of agency costs in resolution therefore is positive. Courts and commentators in other contexts reach a similar conclusion. In corporate governance, the business judgment rule protects the board of directors from liability for corporate decisions not infected by illegality or conflict of interest. The rule prevents a court from vetting a board’s decision even when the decision harmed shareholders’ interests. By limiting judicial oversight, the rule enables directors sometimes to act carelessly. The rule’s policy justification is that the quality of directorial decisions in the range of cases is worth the price.168 The justification implicitly recognizes that defensible legal rules can create or allow agency costs. The observation applies equally to the FDIC’s decisions about asset dispositions and the resolution process. Thus, the FDIC’s control of the resolution process is not suspect simply because it creates agency costs in asset dispositions.

Most important, the arguments above take the existing capital structure of banks as given. The structure is one in which insured deposits predominate; banks have some secured debt, and rarely, much subordinated or general debt. As a practical matter the FDIC owns failed banks: It has the residual claim on the banks’ assets. Given this structure, it is hard to find other actors with better incentives, even if the incentives of the FDIC’s employees are imperfect. Insured depositors have no incentive to see that the sale of the bank’s assets gets top dollar: Their deposits are protected; however, the assets are liquidated. As noted above, the senior claimants (FHLB loans and other secured loans) and junior claimants (general creditors, subordinated debt holders and equity holders) obviously do not have the right incentives, as the former are always paid in full and the latter will almost always receive nothing. One could grant some decision-making power to the very few uninsured depositors who have failed to get out before failure and

168. See, e.g., Stephen M. Bainbridge, The Business Judgment Rule as Abstention Doctrine, 57 VAND. L. REV. 83, 109 (2004) ("In the director primacy model . . . the business judgment rule is justified precisely because judicial review threatens the board’s authority.").

1030 67 WASH. & LEE L. REV. 985 (2010)

perhaps have a judge adjudicate disputes among them. However, this supervision would impose its own costs. As noted in Section III.B., most uninsured depositors withdraw their funds prior to the bank’s failure, and the depositors that remain behind will, almost by definition, be the least attentive. These depositors are likely to be widely dispersed, and they therefore may be poorly situated to monitor the resolution process. Recall that the push for a centralized agency with power to resolve failed banks stemmed not from deposit insurance but from a perception that existing receiverships were plagued by inefficiency and self-dealing by receivers.

Although we lack direct measures of these agency costs, the literature offers some related evidence. A study by Christopher James shows that (1) the FDIC’s losses tend to increase with the percentage of the bank’s assets that it retains, and (2) the direct costs of bank insolvencies average around 10% of assets, far higher than the averages found in other studies of bankruptcy reorganizations.169 Neither result clearly establishes the presence of substantial agency costs. The first result is consistent with the FDIC not being as adept as private banks at selling or managing bank assets. It is also, however, consistent with the possibility that acquirers are more willing to purchase low-risk assets or performing loans due to adverse selection. James tries to distinguish between these possibilities. Although he finds no evidence that acquirers are less likely to purchase risky assets, he does not disprove this possibility either. The second result is harder to interpret 169. See Christopher James, The Losses Realized in Bank Failures, 46 J. FIN. 1223, 1225, 1228 (1991) (noting a loss of approximately 10% in 1985 through 1986). Bennett and Unal estimate mean direct resolution costs at 3.53% of book value, based on a large sample of bank failures between 1986 and 2007. Rosalind L. Bennett & Haluk Unal, The Cost Effectiveness of the Private-Sector Resolution of Failed Bank Assets, app. Table 2 (FDIC Ctr. for Fin. Research Working Paper No. 2009-11, 2010), http://ssrn.com/abstract=1558445 (last visited Sept. 25, 2010) (on file with the Washington and Lee Law Review). For estimates of resolution costs incurred by the Resolution Trust Corporation in disposing of failed savings and loan banks between 1989 and 1995, see Mike Spaid & Edward D. Thomas, Evolution of the RTC’s Resolution Practices, in FED. DEPOSIT INS. CORP., MANAGING THE CRISIS: THE FDIC AND RTC EXPERIENCE 1980–1994 113, 138–39 (1998). Other estimates which count opportunity costs and declining asset values due to regulatory forbearance as resolution costs are higher. See William P. Osterberg & James J. Thomson, UNDERLYING DETERMINANTS OF CLOSED-BANK RESOLUTION COSTS 6 (1994), available at http://www.clevelandfed.org/research/workpaper/ 1994/wp9403.pdf (noting 21%). By comparison, the direct costs of traditional Chapter 11 reorganizations are much lower. Estimates of median direct costs vary between 1.4% and 1.69% of pre-bankruptcy assets. See Brian L. Betker, The Administrative Costs of Debt Restructurings: Some Recent Evidence, 26:4 FIN. MGMT. 56, 57 (1997) (noting 3.93%); Arturo Bris et al., The Costs of Bankruptcy: Chapter 7 Liquidation Versus Chapter 11 Reorganization, 61 J. FIN. 1253, 1284 (2006) (noting 1.69%); Lynn M LoPucki & Joseph W. Doherty, The Determinants of Professional Fees in Large Bankruptcy Reorganization Cases, 1 J. EMPIRICAL LEGAL STUD. 111, 113 (2004) (noting 1.4%).


because it is unclear if one should expect similar direct bankruptcy costs in bankruptcy and bank insolvency. When calculating direct costs of bankruptcy reorganizations scholars typically include the cost of hiring attorneys and business consultants but do not include the salaries of the managers of the firm. By contrast, once the initial purchase and assumption transaction is complete, the failed bank has no management left, and the FDIC must incur the full costs of managing the estate. In addition, some assets may simply be more expensive to administer than others.

Further evidence of the FDIC’s performance comes from studies of the return to holding the stocks of banks that acquire failed banks. If the FDIC does a poor job of auctioning the failed firm and routinely offers too good of a deal to acquirers, the share price of publicly traded acquirers should rise on the announcement of the P&A agreement. Of course, the acquirers stock may rise as long as the assets (or customer base) of the failed bank are somewhat unique so that potential buyers assign different valuations and the acquirer captures some of the gains from trade. However, a very substantial rise in price would provide some cause for concern. The literature does not, however, conclusively show that the FDIC is selling failed banks for far less than they are worth. Some early studies do find abnormal returns to acquiring banks170 while others do not.171 Part of the divergence in findings may be due to changes in the process of resolving failed banks; many of the studies finding no abnormal return focus on thrifts resolved through public option. Unfortunately, no study looks at transactions since the least cost resolution standard was adopted in 1992.

B. The Liability Structure of Large Banks

Section III shows that the overwhelming majority of failed banks have a similar liability structure. The majority of their debts are deposits, most of 170. See generally William J. Bertin et al., Failed Bank Acquisitions and Successful Bidders’ Returns; 18:2 FIN. MGMT. 93 (1989); Rebel A. Cole et al., Asymmetric Information and Principal-Agent Problems as Sources of Value in FSLIC-Assisted Acquisitions of Insolvent Thrifts, 8 J. FIN. SERVICES RES. 5 (1994); Christopher James & Peggy Weir, An Analysis of FDIC Failed Bank Auctions, 20 J. MONETARY ECON. 141 (1987). 171. See Thomas F. Gosnell et al., The Acquisition of Failing Thrifts: Returns to Acquirers, 22:4 FIN. MGMT. 58, 60 (1993) (finding no abnormal returns); Paul M. Horvitz & Insup Lee, Abnormal Returns in Post-FIRREA Acquisitions of Failed Thrifts, 8 J. FIN. SERVICES RES. 269, 269 (1994) (announcing that "acquisitions of failed S&Ls in the post-FIRREA period have not been accompanied by positive abnormal returns"); Richard H. Pettway & Jack W. Trifts, Do Banks Overbid When Acquiring Failed Banks?, 14:2 FIN. MGMT. 5, 5–6, 12 (1985) (finding negative abnormal returns).

1032 67 WASH. & LEE L. REV. 985 (2010)

which are insured. They also have no or very little foreign deposits, general unsecured debt or subordinated debt. Some banks do have secured liabilities, but these liabilities are always paid in full. This makes the FDIC the likely residual claimant in nearly all bank failures. However, a properly designed bank insolvency law must take into account the likely residual claimant on most of the assets of failed banks. It must consider the size of assets of failed banks, not just the frequency with which banks fail. Showing that the FDIC is the residual claimant in 99% of bank failures means little if the remaining 1% of banks own almost all the assets of failed banks. Asset size matters because the liability structure of the very largest banks differs from that of banks with fewer assets.

In the United States, banking assets are, in fact, concentrated in the largest banks. This remains true whether one examines banks generally or failed banks in particular. In 2008 about 1.4% of insured financial institutions (commercial and savings and loan banks) had assets greater than $10 billion. These "mega-banks" owned about 78% of all assets of insured financial institutions,172 and the asset share of mega-banks has increased consistently over time.173 Measured by their assets, mega-banks also predominate in bank failures. Insured bank institutions with assets greater than $10 billion account for about 3% of the banks that failed between January 1, 1995 and May 31, 2009. However, these banks held 87% of the assets of banks in failure during our sample period.

The capital structures of most failed banks with assets greater than five or ten billion dollars do, in fact, look like those of most other failed banks. Since 1995 the FDIC has resolved five banks with real assets greater than $5 billion (adjusted for inflation). The four smallest of these had deposits that ranged between 64% and 82% of liabilities and insured deposits that ranged between 76% and 91% of total deposits; none of these four banks had any foreign deposits.174 Secured loans (FHLB plus repurchase agreements) accounted for most of the remaining liabilities; these loans ranged from 18% to 35% of total liabilities. None of these four banks had appreciable amounts

172. U.S. CENSUS BUREAU, FDIC-INSURED FINANCIAL INSTITUTIONS BY ASSET SIZE: 2008, available at http://www.census.gov/compendia/statab/2010/tables/10s1142.pdf. 173. Kenneth D. Jones & Tim Critchfield, Consolidation in the U.S. Banking Industry: Is the "Long, Strange Trip" About to End?, 17 FDIC BANKING REV. 31, 34–36 (2005). 174. In the aggregate, banks with more than $10 billion in assets do have significant foreign deposits. Deposits in foreign offices account for about one-quarter of all deposits at these banks. FDIC, Statistics on Depository Institutions Report, http://www2.fdic.gov/sdi/ main.asp (last visited Sept. 25, 2010) (on file with the Washington and Lee Law Review).


of subordinated debt or general claims; the sum of these two categories ranged from 0.9% to 1.9% of total liabilities.

However, the fifth bank, Washington Mutual, had a capital structure that was appreciably different than the rest, and Washington Mutual was far larger than all of the other banks put together. In fact, Washington Mutual Bank alone accounts for more than 70% of the assets of the failed banks in our sample.175 Washington Mutual had significant debts that were junior to the claims of the FDIC. Washington Mutual’s subordinated debt accounted for 2.8% and its general claims 9.9% of liabilities.176 It is dangerous to draw conclusions from a sample of one, but seemingly healthy banks of similar size also have significant claims that are junior to the FDIC. For example, Wells Fargo has subordinated debt that accounts for 3.2% of its liabilities and general claims that account for about 11.6%. Other banks have significant debt that is junior to the FDIC in the form of foreign deposits. Foreign deposits represent about 13% of Bank of America’s liabilities and 68% of Citigroup’s liabilities.177

The presence of significant debt junior to the FDIC’s claim makes it less likely that the FDIC would be the residual claimant should insolvency necessitate a resolution. The least-cost resolution standard requires the FDIC ordinarily to resolve a failed bank in a way that imposes the least cost on the insurance fund.178 The FDIC is not explicitly required to do so in a way that maximizes the return to the junior claims as well. Subrogated to the rights of insured depositors, the FDIC therefore has the incentive only to receive an 175. See supra note 27 (observing that Washington Mutual accounted for $307 billion of $417 billion worth of assets of banks that failed between January 1, 1995 and October 19, 2009). Large nonbank firms represent a fraction of Chapter 11 bankruptcies. In 2008 less than 1% of all bankruptcy cases were initiated by Chapter 11 filings, and only about 0.003% were filed by large public corporations. See Bankruptcy Statistics, http://www.uscourts. gov/Statistics/BankruptcyStatistics.aspx (last visited Sept. 25, 2010) (on file with the Washington and Lee Law Review). According to Lynn LoPucki, of the 1,117,771 bankruptcy filings in 2008, there were thirty-nine large public bankruptcies among 10,160 Chapter 11 filings. LYNN M. LOPUCKI’S BANKR. RESEARCH DATABASE, LARGE, PUBLIC COMPANY BANKRUPTCIES FILED IN THE UNITED STATES, BY YEAR, 1990–2009, available at http://lopucki.law.ucla.edu/Filings%20by%20year.pdf. 176. These figures are taken from Washington Mutual’s final Thrift Supervision report filed before it closed. In the ten weeks before closing, depositors withdrew about 9% of deposits made by the bank, and this would have changed its liability structure. Robin Sidel et al., WaMu is Seized, Sold Off to J.P. Morgan, In Largest Failure in U.S. Banking History, WALL ST. J., Sept. 26, 2008, at A1. 177. See 2009 Call Reports for Bank of America and Citigroup (providing these statistics). 178. See supra note 84 and accompanying text (discussing this requirement imposed by Congress).

1034 67 WASH. & LEE L. REV. 985 (2010)

amount equal to the insured depositors’ claims. For this reason, the FDIC’s control over the resolution process in large bank failures is unlikely to maximize the value of the failed bank’s assets.

Washington Mutual Bank’s resolution illustrates this concern. At its closing, the bank’s assets had a book value of $307 billion. The FDIC transferred most of the assets to J.P. Morgan Chase Bank as part of a purchase and assumption transaction. In return J.P. Morgan paid $1.9 billion and acquired certain of Washington Mutual’s contracts and all of its deposits and liability to general creditors. J.P. Morgan did not assume Washington Mutual’s subordinated debt or equity.179 Subordinated debt and equity will receive nothing while the depositors and general creditors will be paid in full by J.P. Morgan. Because J.P. Morgan acquired all of Washington Mutual’s deposits, there was no loss to the FDIC. If J.P. Morgan paid top dollar for Washington Mutual’s assets, subordinated debt has no complaint. It has a complaint if another bidder would have paid more: a higher bid price might have allowed distributions to subordinated debt. Because J.P. Morgan’s bid resulted in no loss to the insurance fund, the FDIC had no incentive to seek or accept a higher-valued bid. Washington Mutual’s holding company, the holder of subordinated debt, has sued the FDIC, presumably on the theory that the liquidation value of Washington Mutual’s assets is sufficiently higher than the sum J.P. Morgan paid to partly or wholly satisfy its claims.180 Its suit might fail on the merits. However, Washington Mutual’s liability structure does not justify a presumption in favor of the FDIC: Because the FDIC is not the residual claimant, its control of the resolution process cannot be assumed to maximize asset values.

The particular liability structure of large banks does not necessarily undermine the case for FDIC control of the bank resolution process. This exception applies only to the very largest banks, and these failures are very rare—these are the firms that are "too big to fail." The FDIC is the residual claimant in almost all bank failures. At most, the liability structure of large banks justifies special resolution rules for large banks or banks with substantial claims subordinate or equal to the FDIC in priority that would constrain the FDIC’s discretion or force it to share control. There is some precedent for special rules tied to the size or nature of the debtor. The

179. Purchase and Assumption Agreement Among Federal Deposit Insurance Corporation and J.P. Morgan Chase Bank, Sept. 25, 2008, Article 2.1, Schedule 2.1, at 8, 34, available at http://www.fdic.gov/about/freedom/Washington_Mutual_P_and_A.pdf. 180. See Washington Mutual Inc. v. Federal Deposit Insurance Corp., Civ. Action No. 09-533(RMC), 2009 WL 3273880, at *1 (D.D.C. 2009) (considering a motion by note holders to intervene as a defendant).


Bankruptcy Code includes separate provisions for small business debtors in Chapter 11 by way of separate reporting requirements and deadlines.181 Their underlying rationale apparently is that small businesses seeking to be reorganized differ enough from larger businesses to warrant some special rules. In large bank insolvencies, special rules might better protect other creditors as it is less likely that the FDIC would be the only residual claimant. These rules could include enhanced bid procedures, input by junior debt, or an increase in the potential liability of the FDIC to junior debt in connection with the disposal of assets.182

We do not advocate such rules because banks may manipulate their assets or liabilities for reasons unrelated or antithetical to the purpose behind these rules.183 By offering higher interest rates a bank could raise deposits in the short run and use the funds to acquire the assets necessary to trigger the special rules.184 Alternatively, the bank could obtain equity injections 181. See, e.g., 11 U.S.C. §§ 101(51D), 308(b), 586(a)(7), 1121(c), 1125(f) (2006) (providing procedures for small business debtors). 182. Recent legislative proposals move in the opposite direction. As we discuss in the next section, these proposals would place the FDIC in charge of resolving large bank holding companies even when it has no claim on the assets of these firms. 183. A separate set of legal rules for specific sorts of actors by itself is not unusual or controversial. Commercial and securities law sometimes has separate rules tailored to different sorts of transactions or transacting parties. Perhaps the best illustration is Article 2 of the Uniform Commercial Code’s rules governing merchants. See, e.g., U.C.C. § 2-104(1), 2-205, 2-207(2)(a), 2-314(a), 2-509(2) (2009) (providing various rules that apply to those considered "merchants"). In a securities offering Regulation D gives a safe harbor based on the size of the offering: limited offerings below $1–5 million are not subject to enhanced disclosure requirements; offerings above $5 million are public offerings requiring mandated disclosures. 17 C.F.R. § 230.504–505 (2009). Another safe harbor allows issuers to raise unlimited amounts of money from accredited investors in private offerings. See id. § 230.506 (providing such an exemption for limited offerings). These safe harbors effectively allow securities issuers to select the amount of disclosure they must make by adjusting the amount of their securities offering. However, usually such rules are triggered by attributes of affected parties that do not allow for strategic manipulation by them. Merchant status, for instance, is an attribute that cannot be manipulated in the short term by a seller or buyer of goods. Accredited investor status also is unalterable in the short term. By contrast, banks can manage their assets or liabilities in the short-term, particularly through transactions with their affiliates or holding companies, to achieve a desired size. 184. There are numerous examples of troubled banks using high interest rates to attract deposits. See, e.g., Kathleen Pender, Tricky Situations over Interest Rates at Ally Bank, S.F. CHRON, June 16, 2009, at C1 ("Although many well-rated banks offer above-average yields, it’s not unusual to see banks on the verge of failure offering the highest rates in an effort to stay afloat."). "IndyMac Bank was at or near the top of Bankrate’s list before it was taken over by the FDIC last summer. Countrywide was also one of the top yielders before it was forced into Bank of America’s arms." Id. The FDIC is aware of this risk and limits the interest rate that banks can offer when they are deemed to be not well capitalized. See 12 C.F.R. § 337 (taking effect December 3, 2009, banks that are not well-capitalized cannot

1036 67 WASH. & LEE L. REV. 985 (2010)

disguised as subordinated debt from its holding company. In the final period before insolvency the depositors may flee, the subordinated debt may disappear, and the assets may fall in value. However, these changes in the financial condition might not be reflected in the bank’s quarterly Call Report before failure. As a result, the resolution authority may be unable to tell how control should be allocated until it audits the bank’s financial statements. This takes time. Moreover, if the special rules are triggered by size, the bank could qualify as "large" without its creditors having a residual interest in its assets.

The prospect of recapitalization or other financial support unavailable in the alternative bank resolution procedure creates moral hazard. The availability of financial support may confer a competitive advantage on banks deemed to present systemic risk. This could induce further bank consolidation that in turn increases such risk in the first place. We do not argue that a separate resolution procedure for large banks is unjustified under all conditions. Our point is more modest: the unique liability structure of the largest banks does not necessarily undermine the case for the FDIC’s control over the resolution of failed banks.

C. Self-Financed Restructuring: Debt Conversion

By asking whether the FDIC should control the resolution of insolvent banks, this Article presumes the need of a resolution procedure. A number of authors have suggested regulatory reforms designed to resolve insolvency without a formal procedure.185 Specifically, they suggest that banks be forced to issue substantial amounts of debt that would convert into equity as the value of the firm’s assets declines.186 We do not engage this debate for two

offer a rate higher than seventy-five basis points above the rate prevailing in the market). The prospect of a favorable resolution procedure exacerbates the incentive of troubled banks to attract deposits. 185. See, e.g., Mark J. Flannery, No Pain, No Gain? Effecting Market Discipline via "Reverse Convertible Debentures," in CAPITAL ADEQUACY BEYOND BASEL: BANKING, SECURITIES, AND INSURANCE 171, 171 (Hal S. Scott ed., 2005) (proposing "a new financial instrument that forestalls financial distress without distorting bank shareholders’ risk-taking incentives"); Squam Lake Working Group on Financial Regulation, An Expedited Resolution Mechanism for Distressed Financial Firms: Regulatory Hybrid Securities 2 (Council on Foreign Relations Working Paper, Apr. 2009) (recommending "support for a new regulatory hybrid security that will expedite the recapitalization of banks"). 186. Consider an example based on Flannery’s article. Assume that a bank begins with assets of $1,000, must maintain equity capital equal to 8% of its assets ($80), and has convertible debt equal to an additional 5% of its assets ($50). If the bank’s assets decline in


reasons. First, we take the existing capital structure of banks as given. Second, convertible debt (or enhanced equity requirements) can reduce, but cannot eliminate, the risk of insolvency. If the bank’s assets fall sharply in value, there may not be enough convertible debt to return the bank to solvency.

Debt conversion can, however, be used as a resolution mechanism. Chapter 11 of the bankruptcy code is, in fact, one such mechanism. The earliest bankruptcy systems resolved failed firms by selling their assets in the market and distributing the proceeds to their creditors in accordance with priority.187 This is still the predominant method used today; most bankrupt firms are liquidated in Chapter 7.188 The FDIC’s favored resolution method, a purchase and assumption, in effect is a liquidation coupled with the assumption of some of the failed bank’s liabilities by another firm.189 By contrast, a traditional Chapter 11 reorganization can resolve a failed firm without an actual sale of its assets.190 The assets are instead effectively transferred among the existing claimants in a hypothetical sale. Junior claims are cancelled if the assets are insufficient to pay the more senior claimants in full, and some of the remaining debt is typically converted into equity to return the firm to solvency. The value of the firm’s assets is determined by

value to $970, capital requirements would force the bank to have $77.50 in equity, and so $27.60 of the debt would convert into equity. If the bank’s assets fall further in value to $900 (far enough to render the firm insolvent if the $50 in initial debt had not been convertible), the remainder of the convertible debt would be exchanged for equity. The bank would eventually need to issue more equity, but it would remain solvent. If asset prices fall sufficiently slowly to allow regulators to force the bank to continually issue more convertible debt securities, this system could ensure that the bank never becomes insolvent. Note, however, that regulators could also keep the bank solvent by simply setting equity capital requirements at 13% instead of setting an equity capital requirement of 8% and requiring an additional 5% of convertible debt. Convertible debt may, however, offer other advantages. For example, the ability of the bank to issue junior debt and the interest rate that it must pay may provide useful information to regulators about the true risk of the bank’s assets. These pre-insolvency advantages are beyond the scope of this Article. 187. See, e.g., Bankruptcy Act of 1800, ch. 19, §§ 29–31, 2 Stat. 19, 29–30 (1800) (providing for division of bankrupt’s estate); Bankruptcy Act of 1841, ch. 9, §§ 3–4, 5 Stat. 440, 442–444 (providing for seizure of bankrupt’s property). 188. See 11 U.S.C. §§ 701–84 (2006) (providing rules for liquidation process in bankruptcy); Bankruptcy Statistics, United States Courts, http://www.uscourts.gov/ bnkrpctystats/statistics.htm#june (last visited Sept. 25, 2010) (providing bankruptcy statistics) (on file with the Washington and Lee Law Review). 189. See supra notes 11–15 and accompanying text (discussing FDIC resolution of bankruptcies). 190. See supra notes 40–52 and accompanying text (discussing Chapter 11 bankruptcy).

1038 67 WASH. & LEE L. REV. 985 (2010)

the agreement of the parties in a consensual reorganization or by the bankruptcy judge in a cramdown.

Recently, at least one scholar has drawn upon prior literature to propose a bank resolution mechanism that would convert a troubled bank’s capital structure "automatically" or with minimal proceedings.191 This conversion mechanism would follow the priority structure so that junior claims could be cancelled if the assets of the bank were less than the aggregate value of more senior claims.192 For example, assume that banks must meet a required minimum capital ratio of 8%. Suppose a bank that has issued 10 shares of common stock becomes insolvent as the value of its assets fall from $1,000 to just $800. Before resolution, its balance sheet reveals:

Assets Liabilities $800 $870 Deposits $50 General claims

($120) Equity (10 shares)

The bank’s assets are sufficient to satisfy only $800 of the $920 in outstanding claims against it. Thus, to return to solvency, the ten shares and $120 in debt must be cancelled. To meet the minimum capital requirements an additional $64 of debt must be converted into equity. Because general claims have lower priority than deposits, the general claims totaling $50 are extinguished. Similarly, $70 of the deposits must be extinguished and $64 of the deposits must be converted into equity,193 leaving $736 in deposit 191. See Luigi Zingales, Yes We Can, Secretary Geithner, 6 ECONOMIST’S VOICE 1, 3 (2009) [hereinafter Zingales, Yes We Can] (proposing "a new form of bankruptcy for banks, where derivative contracts are kept in place and the long term debt is swapped into equity"). Professor Zingales does not fully develop this mechanism in his short piece, Luigi Zingales, Plan B, 5 ECONOMIST’S VOICE 1 (2008) [hereinafter Zingales, Plan B] (discussing solutions to the financial crisis). For earlier proposals that apply to all firms, not just banks, see Barry E. Adler, Financial and Political Theories of American Corporate Bankruptcy, 45 STAN. L. REV. 311, 323 (1993) (proposing a "Chameleon Equity" approach); Note, Distress-Contingent Convertible Bonds: A Proposed Solution to the Excess Debt Problem, 104 HARV. L. REV. 1857, 1869–77 (1991) (discussing distress-contingent convertible bonds). Many of the proposals call for debt conversion by contract; the description in the text provides for mandatory conversion under the prescribed conditions. See Zingales, Yes We Can, supra note 191, at 3 (calling for legislation creating "a new form of bankruptcy for banks"). 192. See Zingales, Yes We Can, supra note 191, at 4 (discussing the proposal’s effect on junior claimants). 193. The mechanism need not formally cancel any debt that is equal in priority to the residual claim. If, for example, the mechanism "converted" $13.40 in deposits to equity and cancelled no deposits, the market would quickly cancel $7 by revealing that the equity is worth just $6.40.


liabilities. This does not deprive insured depositors of access to the insured balances in their deposit accounts. Insured depositors are reimbursed by the FDIC from the bank insurance fund in the amount of their insured deposits. Their interests therefore are unaffected by the conversion of their debt into equity.194 The FDIC in turn is subrogated to the rights of depositors it reimburses, to the extent of the reimbursement.195 Applicable priority rules give insured and uninsured depositors equal priority in a bank resolution.196 Thus, the FDIC and the bank’s uninsured depositors (if any) are issued shares in proportion to their share of the $64 in deposits converted to equity. The FDIC can sell its shares, as it can sell almost any asset in a bank’s resolution.197

Automatic debt conversion offers speed and reduced administrative costs.198 However, even a truly automatic debt conversion mechanism would likely fail to yield these benefits unless the capital structures of banks are radically changed. If applied to the banks that have actually failed, these proposals would have granted the FDIC most (and in most cases nearly all) of the shares of the failed bank. The FDIC will have to sell these shares; financial constraints force it to do so. The costs of such sales might equal or

194. See Zingales, Yes We Can, supra note 191, at 5 (proposing such a debt conversion scheme). Zingales’ proposed debt conversion scheme for systemically important banks exempts individual depositor liability. See Zingales, Plan B, supra note 191, at 4 ("My plan would exempt individual depositors, who are federally ensured."). Because conversion does not affect the liquidity to insured individual deposits, the exemption is unnecessary. 195. See 12 U.S.C. § 1821(g) (2006) (providing for subrogation). 196. See id. § 1821(d)(11)(A)(ii) (giving equal priority to "[a]ny deposit liability of the institution"). 197. See supra notes 11–17 and accompanying text (discussing resolution by the FDIC). 198. See supra note 169 (discussing costs of resolution). Estimates of the direct costs to the FDIC, although imprecise, are as high as 21% of asset values. Supra note 169. By comparison, the direct costs of traditional Chapter 11 reorganizations are much lower. Estimates of median direct costs vary between 1.4% and 1.69% of pre-bankruptcy assets. See Arturo Bris et al., The Costs of Bankruptcy: Chapter 7 Versus Chapter 11 Reorganization, 61 J. FIN. 1253, 1277, 1284 (2006) (providing a median direct cost of 1.69% of pre-bankruptcy assets); Lynn M LoPucki & Joseph W. Doherty, The Determinants of Professional Fees in Large Bankruptcy Reorganization Cases, 1 J. EMP. LEGAL STUD. 111, 127 (2004) (providing a median direct cost of 1.4% of pre-bankruptcy assets); Brian L. Betker, The Administrative Costs of Debt Restructurings: Some Recent Evidence, 26 FIN. MGMT. 56, 61 (1996) (providing a median direct cost of 3.37% of pre-bankruptcy assets); Lawrence A. Weiss, Bankruptcy Resolution: Direct Costs and Violation of Priority of Claims, 27 J. FIN. ECON. 285, 299 (1990) (providing a median direct cost of 2.5% of pre-bankruptcy assets).

1040 67 WASH. & LEE L. REV. 985 (2010)

exceed the costs the FDIC currently incurs in selling the assets of failed bank through a receivership.199

A debt exchange in almost every past bank failure would have left the FDIC as the majority shareholder because almost all failed banks have no or little debt subordinate to deposits and little uninsured deposits at the time of failure.200 At the same time, insured deposits are the major liabilities of most banks.201 Because the FDIC is subrogated to the claims of insured depositors, it ultimately shares loss with uninsured depositors when a bank fails. The consistent pattern of FDIC loss shows that depositors lose money in most bank failures.202 Thus, a debt exchange will give depositors equity in the typical failed bank. The FDIC will receive the largest share of equity, because insured deposits represent most deposit liabilities of typical failed banks. However, the FDIC is unlike holders of swapped debt in nonbank firms. Creditors in nonbank firms whose debt has been converted to equity might want to retain it. The conversion can give them a control premium in the firm stock. Alternatively, the equity can give a creditor a strategic advantage in the relevant industry. More generally, financial constraints do not generally restrict their ability to hold firm stock. The FDIC is different. It cannot retain large equity positions that come from debt conversions of failed banks. To avoid exhausting the balance of the deposit insurance fund, the FDIC must sell bank assets, and FDIC policy requires as much.203 For the 199. See supra notes 11–15 (discussing receivership under FDIC). 200. Required increases in the amount of unsecured claims or subordinated debt would increase the amount of debt junior to deposit liabilities that could be converted into equity. Alternatively, required increases in the amount of equity would avoid the risk of insolvency altogether. The relative merits of these proposed requirements are beyond the scope of this Article. For assessments of the proposals, see generally Douglas D. Evanoff & Larry D. Wall, Subordinated Debt and Bank Capital Reform (Fed. Res. Bank of Chicago, WP 2000-07, Aug. 2000); Study Group on Subordinated Notes and Debentures, Staff Study 172: Using Subordinated Debt as an Instrument of Market Discipline, (Fed. Res. Sys., Dec. 1999). 201. See supra Part II.B ("Domestic deposits account for the overwhelming majority of bank liabilities, and almost all of these domestic deposits are insured."). 202. See supra Part II.B (discussing losses when banks fail). 203. See supra note 13 and accompanying text (discussing FDIC Asset Disposition Manual suggested timelines for sales). The Federal Deposit Insurance Act requires the balance of deposit insurance fund to satisfy a statutorily mandated minimum "designated reserve ratio": The ratio of the balance of the deposit insurance fund to insured deposits. See 12 U.S.C. § 1817(b)(3) (2006) (discussing the designated reserve ratio). Paying losses to insured deposits reduces the balance in the fund, thereby also reducing the effective reserve ratio. The designated reserve ratio therefore limits the amount available to the FDIC to pay insured depositors. By increasing the fund balance, the FDIC in principle can pay insured deposits while staying within the designated reserve ratio. The fund balance can be increased by increasing assessments against member banks, special assessments, or


same reason, it would have to sell its equity in failed banks obtained under a debt conversion scheme. Thus, debt conversion schemes do not avoid the cost of asset sales in the case of failed banks. They merely postpone those costs so that they are incurred later, outside of the formal resolution process. Again, a plausible accounting must count these costs as resolution costs. These costs might equal the administrative costs the FDIC currently incurs in disposing of failed bank assets.

Automatic debt conversion increases the speed and reduces the cost of the final resolution only if the FDIC can more easily dispose of the stock of a failed bank than it can dispose of the assets of a failed bank through a receivership. We see little reason why this would be true. Shares of a firm are just a claim against all of the assets of that firm. Acquirers of failed banks have demonstrated that they do not want to buy the troubled assets of the failed bank: Most purchase less than one quarter of the failed bank’s assets.204 True, acquirers may be more willing to purchase shares than troubled assets because of the option inherent in a highly leveraged firm. The acquirer would risk only the amount that they paid for the equity. Any further decline in the value of the assets of the firm would be borne by the holders of the debt (mainly the FDIC). However, the FDIC can, and does, effectively replicate this structure in some current transactions. The FDIC frequently sells all or substantially all of the assets of the failed bank to the bank that assumes the deposits. In these sales the FDIC simultaneously assumes the risk of a major decline in the value of these assets through a loss sharing agreement.205

As noted in Part III.B, the capital structure of the very largest banks differ markedly from those of nearly all banks that have actually failed, and it is more plausible that an automatic debt conversion mechanism would not leave the FDIC as the dominant shareholder.206 Because this Article uses data from banks that have actually failed, it has little to say about the very largest banks. Only one truly "mega-bank" has failed in the last fifteen years. However, we question whether an automatic conversion mechanism would really offer significant advantages over the current system. There are less and

borrowing from the Treasury. See id. § 1817(b)(5) (describing emergency special assessments); id. § 1824(a) (describing FDIC’s borrowing authority). Political resistance makes these options infeasible on an ongoing basis. In practice the balance in the deposit insurance fund is fixed. Thus, to maintain the designated reserve ratio, the FDIC generally must sell bank assets in order to pay insured deposits. 204. See supra note 81 and accompanying text (providing relevant statistics). 205. See supra Part II.B (discussing loss sharing agreements). 206. See supra Part III.B. (discussing the FDIC as residual claimant).

1042 67 WASH. & LEE L. REV. 985 (2010)

more serious problems of design. The less serious problem is the authority to trigger the conversion of debt to equity and cancellation of existing equity. Stakeholders do not have complete information about the value of the assets and therefore are likely to have different opinions about the bank’s assets’ values. They therefore are unlikely to all agree on the need for recapitalization. The bank’s balance sheet cannot be used as a reliable basis of information, because it computes assets and equity at book value. Book values tend to lag behind the market value of assets at troubled banks.207 Because stakeholders are unlikely to trigger the debt conversion, the FDIC (or some other entity) must have the authority to trigger it. A plausible and observable trigger is the FDIC’s seizure of a bank. This authority does not give the FDIC control in resolving the bank: Its seizure of the bank simply triggers the conversion of debt to equity. The FDIC’s role is limited to initiating the process by which debt is restructured automatically.

The more serious problem is the design of the financial trigger the FDIC is to use to initiate the resolution process.208 Debt is to be converted to equity when the bank is economically insolvent: When all claims due against the bank cannot be satisfied from its earnings. The conversion mechanism requires some estimate of the market value of the firm’s assets. However, these estimates are sometimes difficult to gauge, indeterminate, and manipulable. Loans are important assets of banks, and nonstandard terms in commercial or installment loans and unobservable risk characteristics of borrowers make many loans difficult to value when not traded in active markets. Something similar might be true of securitized assets on a bank’s books. Market value in both cases is hard to determine. Financial triggers based on share price cannot be used at all for banks in which equity is privately held. Although the market price of traded shares can be used, variance in daily trading prices can be random. Averaging traded prices over a longer period avoids this trouble but brings another: Strategic manipulation of share prices. For instance, debt holders expecting FDIC intervention might short shares of the bank in an effort to drive the price down. A lower share price allows them more shares if their debt is converted to equity.209

207. See Joe Peek & Eric S. Rosengren, The Use of Capital Ratios to Trigger Intervention in Problem Banks: Too Little, Too Late, NEW ENGLAND ECON. REV.49, 51–52 (Sept./Oct. 1996) (noting a lag of reductions in leverage ratios behind deterioration in bank health). 208. See infra notes 220–22 and accompanying text (discussing the resolution process). 209. See Flannery, supra note 185, at 186 (discussing how a trigger ratio should be computed and how often it should be evaluated); Squam Lake Working Group on Financial Regulation, supra note 185, at 4 (discussing triggers).


Although this manipulation is unlikely to be successful with large banks whose stock is actively traded, it might work with smaller banks whose shares are less actively traded. A financial trigger geared to share value also might entice management to try to lower share price by bringing FDIC intervention, for the purpose of having conversion result in more shares issued.

These problems in design do not make a debt conversion scheme unworkable. However, they make implementing it infeasible for bank resolutions. Methods of estimating the bank’s value that use market mechanisms likely cannot be applied. They rely on time and liquidity that may not be available in the case of many failed banks. Take Bebchuck’s proposal to give stakeholders an option to purchase senior claims for their face amount.210 Assume that the bank has two creditors, each with $100 claims having equal priority. Suppose the market price of the bank is $200. Debt conversion cancels the shareholder’s stock and transforms the creditors’ $200 claims to equity. Under Bebchuck’s proposal, the shareholders are entitled to regain their stock for the face amount of claims.211 If shareholders believe that the bank is worth more than $200, they will repurchase their stock. If they believe the bank is worth less than $200, they will not. In this way the repurchase option avoids the need to rely on market price in converting debt to equity.

Bebchuck’s proposal requires liquidity and time that often is missing in bank failures. The proposal sets a stipulated short period in which stakeholders can exercise their options.212 In the case of failed banks, this period must be very short. Otherwise, uncertainty about the bank’s capital structure risks disrupting its operations. To exercise their options within the period, stakeholders need funds quickly. This usually is not a problem where the price of exercising their options is small relative to their wealth. But liquidity is a problem when the required exercise price is proportionately large even for institutional shareholders. Stakeholders with large illiquid holdings can borrow funds in amounts needed to exercise their options. However, the ability to borrow quickly is not guaranteed. A lender will not make an unsecured loan for $200 if the market price of the shares in a bank is less than $200. Although it might lend $200 on a secured basis, the lender 210. See Lucian Arye Bebchuk, A New Approach to Corporate Reorganizations, 101 HARV. L. REV. 775, 785 (1988) (proposing a new "method of dividing the reorganization pie"). 211. See id. at 781 (providing an example of the proposed approach). 212. See id. at 789 (proposing a "brief period between the issue date and the exercise date").

1044 67 WASH. & LEE L. REV. 985 (2010)

needs time to value collateral and obtain the required security interest. The time needed can exceed the period in which the stakeholder can exercise its option. Because the stipulated period must be very short in bank failures, liquidity problems likely prevent stakeholders from exercising their purchase options.213 In these cases the market price of the bank effectively is final: It determines the division of value among the bank’s stakeholders according to their priority. By comparison, the pricing mechanism used in FDIC-led bank resolutions works quickly. The FDIC solicits bids prior to closing the bank; the winning bid usually determines the value of the bank.

V. Bank Holding Companies and Nonbank Financial Companies

So far this Article has focused on the control of the resolution of failed banks. We turn now to the holding companies that own the stock of these banks or firms offering financial services. Most banks are affiliates of bank holding companies.214 While their bank subsidiaries are resolved by the FDIC, the bank holding companies and their nonbank subsidiaries can, and do, file for protection under the Bankruptcy Code.215 Part III argued that the FDIC should control the disposition of a failed bank’s assets because it is the residual claimant on those assets.216 This section asks whether this justification supports proposals to alter the resolution procedure for certain bank and financial holding companies. We conclude that it does not.

The comprehensive financial reform bill recently enacted by Congress changes this allocation of resolution authority for the largest bank holding companies and nonbank financial companies.217 The legislation creates an

213. See Philippe Aghion et al., The Economics of Bankruptcy Reform, 8 J. L. ECON. & ORG. 523, 540 (1992) (providing alternatives to Chapter 11 reorganization). The alternative proposed by Aghion et al. also gives stakeholders purchase options on equity. See id. (discussing this proposal). Their stipulated period of one month in which these options are exercisable enhances the ability of stakeholders to obtain needed funds. See id. at 535 (proposing a one month exercise period). However, a month is far longer than the time in which bank failures plausibly must be resolved. 214. See supra note 29 and accompanying text (noting the high proportion of U.S. banks owned by bank holding companies). 215. Some of their nonbank subsidiaries may also be ineligible for bankruptcy protection. For example, insurance companies cannot file for bankruptcy. See 11 U.S.C. § 109(b)(2) (2006) (listing exclusions to the definition of debtor under chapter 7). 216. See supra Part III (discussing FDIC control of bank insolvencies). 217. See Dodd-Frank Wall Street Reform and Consumer Protection (Dodd-Frank) Act, Pub. L. No. 111-203 (West, Westlaw through 2010 legislation), H.R. 4173, 111th Cong. (2010) (to be codified in scattered sections of 12 U.S.C.).


"orderly liquidation authority" (OLA) for "covered financial companies."218 Covered financial companies essentially include bank holding companies, their nonbank and broker dealer subsidiaries, and nonbank financial companies whose revenues mostly derive from financial activities.219 These companies are subject to the OLA’s special resolution procedure if their actual or likely failure would have "serious adverse effects on financial stability in the United States."220 Although undefined, these effects presumably include systemic risk to the U.S. financial market created by default. The procedure is triggered by a voting among designated authorities according to a voting rule adapted from the FDIC voting rule required for a finding of systemic risk.221 Once triggered, the Secretary of Treasury will appoint the FDIC as receiver to liquidate the financial company.222 (The FDIC in turn will appoint the Securities Investor Protection Corporation as receiver to liquidate registered brokers or dealers subject to the OLA.)223 Initiation of the Resolution Act’s resolution procedure displaces the Bankruptcy Code and ends an ongoing bankruptcy case of the holding company or its covered subsidiaries.224 The OLA requires the FDIC to dispose of assets in a manner which, inter alia, is "necessary for purposes of the financial stability of the United States and not for the purpose of preserving the covered financial company."225

The OLA’s special resolution procedure gives the FDIC the authority to provide direct assistance to these systematically important covered companies, including covered bank holding companies.226 The assistance

218. See id. §§ 201(a)(11), 202(a)(1), H.R. 4173 §§ 201(a)(11), 202(a)(1) (defining the term "financial company" and instructing how to properly commence an orderly liquidation). 219. See id. § 201(a)(11), (b), H.R. 4173 § 201(a)(11), (b) (defining the term "financial company" and outlining the definitional criteria). 220. Id. § 203(b)(1), (2), H.R. 4173 § 203(b)(1), (2). 221. See id. § 203(a)(1), H.R. 4173 § 203(a)(1) (requiring parties to vote with regards to the written recommendation); cf. 12 U.S.C. § 1823(c)(G) (2006) (providing for emergency determination of systemic risk by the FDIC Board of Directors); Financial Stability Improvement Act of 2009, H.R. 3996, 111th Cong. § 1603(a)(1) (2009) (providing a similar voting procedure for the Federal Reserve Board). 222. See Dodd-Frank § 204(b), H.R. 4173 § 204(b) (appointing the FDIC as receiver). 223. See id. § 205(a), H.R. 4173 § 205(a) (instructing the FDIC to appoint the Securities Investor Protection Corporation as a trustee for the liquidation). 224. See id. § 208(a), H.R. 4173 § 208(a) (dismissing any case or proceeding commenced under the Bankruptcy Code). 225. Id. § 206, H.R. 4173 § 206. 226. See id. § 204(d), H.R. 4173 § 204(d) (authorizing the OLA to make loans to the covered financial company, purchase or guarantee the assets of the covered financial company, or assume or guarantee the obligations of the covered financial company); id.

1046 67 WASH. & LEE L. REV. 985 (2010)

comes from sums borrowed from the Treasury Department.227 To repay Treasury, the OLA establishes a segregated fund in the Treasury (the Orderly Liquidation Fund, or OLF) supplied by assessments on covered financial companies.228 Commentators debate whether failing bank holding companies should receive government assistance.229 One point of dispute is systemic risk. Systemic risk, although often invoked, is hard to define and arguably regulate effectively. The failure of financial institutions can produce significant loss in financial institutions or nonfinancial markets. However, the distinction between cascading losses resulting from a financial company’s failure and large macroeconomic effects produced by its failure is elusive. Significantly, the OLA employs but does not define the operative notion of systemic risk.230 Another dispute is over the consequences of financial assistance. Assistance may dampen the effects of financial breakdowns and thus avert major losses to the real economy. On the other hand, assistance represents a cross-subsidy from contributing holding companies which do not create systemic risk. The cross-subsidy makes financial institutions more willing to engage in risky behavior and may make financial breakdowns more likely.

This Article takes no position on systemic risk or how effectively the OLA or supervisory authority regulates it.231 It focuses instead on control of the § 210(a), H.R. 4173 § 210(a) (giving the FDIC a wide range of enumerated powers to assist covered financial companies). In exercising these powers the OLA is instructed to ensure that the "creditors and shareholders will bear the losses of the financial company." See id. § 204(a), H.R. 4173 § 204(a) (discussing the purpose of OLA). 227. See id. § 210(n)(5), H.R. 4173 § 210(n)(5) (listing the parties that have the authority to issue obligations). 228. See id. § 210(n)(1), (o)(1)(B), H.R. 4173 § 210(n)(1), (o)(1)(B) (establishing the OLF and requiring the FDIC to recover risk-based assessments on certain financial companies when necessary to repay the Treasury). 229. See, e.g., Ken Ayotte & David A. Skeel, Jr., Bankruptcy or Bailouts?, 35 J. CORP. L. 469, 483–88 (2010) (discussing the issues surrounding government assistance to failing banks). 230. See Dodd-Frank § 203(b)(2), H.R. 4173 § 203(b)(2) (requiring a determination that the failure would have "serious adverse effects on the financial stability of the United States"); see also George G. Kaufman & Kenneth E. Scott, What is Systemic Risk, and Do Bank Regulators Retard or Contribute to It?, 7 INDEP. REV. 371, 372–75 (2003) (providing a description of different notions of systematic risk). 231. We also do not address two subsidiary questions. One is the choice of governmental agency to provide assistance. The second question concerns how the assistance is funded. In the recent crisis Treasury provided money to bank holding companies pursuant to the Troubled Asset Relief Program. See generally U.S. DEP’T OF TREASURY, TARP CAPITAL PURCHASE PROGRAM, available at http://www.ustreas.gov/press/ releases/reports/document5hp1207.pdf. Some of the motivation for the Resolution Act is a desire to shift the funding of this assistance from general tax revenue to a special fund administered by the FDIC and raised by assessments on large bank holding companies. This


resolution process itself. To begin, notice that FDIC-provided financial assistance does not by itself require or justify FDIC control over a bank holding company’s resolution. Such assistance gives the FDIC a claim against the bank holding company in its resolution. But this says nothing about whether the FDIC must or should control the company’s resolution process. For example, the federal government provided substantial loans and other assistance to General Motors both before and after it filed for bankruptcy and thus had a claim on General Motors’ assets.232 Although the government may have played a major role in the General Motors bankruptcy, the other claimants were at least entitled to notice and a hearing before General Motors sold most of its assets.233 Similarly, if the Treasury or Federal Reserve234 provides assistance to a bank holding company it will have a claim against the company. Under present law, in the holding company’s bankruptcy, the government shares control with all other claimants against the holding company’s estate. It does not completely control the holding company’s resolution. In short, current law adopts a rule of "claim without total control." Of course, a change in law could give the FDIC or other government agency total control over the resolution process, even if it had no claim against the holding company. "Total control without a claim" has a historical precedent. A predecessor bill to the 1933 Bank Act proposed a federal agency to dispose of failed banks.235 The proposal was based on perceived inefficiencies in state bank receiverships. Similarly, inefficiencies in the Bankruptcy Code might justify a government-controlled receivership process for bank holding companies. Our point is simply that a claim against a holding company does not by itself require or justify control of the company’s resolution.

The residual claimant principle defended in Part II does not justify the FDIC’s control of the resolution of bank holding companies. The FDIC could have a significant claim against the failed bank holding company, either because it provided direct assistance to the holding company236 or if the bank holding

shift may make direct government assistance more politically palatable by blunting populist anger over the use of general tax revenue to rescue large banks. This shift may also have desirable tax incidence implications, although we know of no study that has conducted the necessary analysis. 232. See supra note 53 and accompanying text (detailing the sales of both Chrysler and GM). 233. Supra note 53 and accompanying text. 234. See 12 U.S.C. § 343 (2006) (providing limited authority for the Federal Reserve Bank to lend to nonbanks). 235. See supra notes 150–51 and accompanying text (describing the proposed Federal Liquidating Corporation, which would have had power to completely liquidate a failed bank regardless of the government’s lack of a claim against the bank). 236. The OLA permits the FDIC to fund companies from sums borrowed from the Treasury

1048 67 WASH. & LEE L. REV. 985 (2010)

company issued a source-of-strength guarantee in connection with a previous capital restoration plan.237 However, we doubt that this claim would often make the FDIC the residual claimant of the systematically important failed bank holding company. The Resolution Act gives the FDIC priority over general claims and subordinated debt238 just as current law grants the FDIC priority over the general claims and subordinated debt of banks.239 Unlike most banks, however, the largest bank holding companies have substantial amounts of general claims and subordinated debt.240 Thus, the FDIC would be the residual claimant only if the

Department. 237. See 12 U.S.C. § 1831(e)(2)(E) (providing that FDIC approval of a capital restoration plan requires the controlling company to guarantee the lesser of 5% of undercapitalized bank’s assets or the amount needed to adequately capitalize bank). The Federal Reserve Board has its own broader source-of-strength authority. See 12 C.F.R. § 225.4(a)(2) (2009) (granting the Board the power to require a bank holding company to cease any financially risky activity). While exercise of that authority can give the Board a claim against the parent, it does not give the FDIC a claim. See id. (granting authority only to the Federal Reserve Board). 238. See Dodd-Frank § 210(b)(1), H.R. 4173 § 210(b)(1) (ordering the priority of claims); cf. Establishing a Framework for Systemic Risk Regulation: Hearing Before the H. Comm. on Banking, Housing, and Urban Affairs, 111th Cong. 7 (2009) (statement of Sheila C. Bair, Chairman FDIC) ("The new resolution powers should result in the shareholders and unsecured creditors taking losses prior to the government . . . ."). 239. See 12 U.S.C. § 1821(d)(11)(A) (providing the FDIC’s position in order of priority); id. § 1813(l)(5) (defining the term "deposit"). 240. Most of the assets of the five largest U.S. bank holding companies are in the form of stock or debt held in their bank and nonbank subsidiaries, as the following chart shows.

Holding Company: March 31, 2009 Unconsolidated Balance Sheet (Selected Items)

J.P. Morgan Citigroup Bank of America

Wells Fargo

Goldman Sachs

Total Assets 444.5 361.5 446.8 257.5 293 Stock in Subs 184.2 (41.4%) 192.6 (53.2%) 241.3 (54%) 133.1 (51.6%) 66.4 (22.6%) Loans to Subs 148.6 (33.4%) 134.1 (37%) 53.2 (11.9%) 78.2 (30.3%) 177.5 (60.5%) Total Liabilities 274.3 217.5 207.3 157.2 229.1

Short Term Debt Commercial Paper

28.5 (10.3%) 0 (0%) 9.1 (4.3%) 5.3 (3.3%) 34.6 (15.1%)

Other 25.3 (9.2%) 10.2 (4.6%) 13.0 (6.2%) 22.3 (14.1%) 155.3 (67.7%) Long Term Debt Subordinated Debt 27.6 (10%) 26.5 (12.1%) 29.5 (14.2%) 13.4 (8.5%) 5.3 (2.3%)

Other Debt 144.5 (52.6%) 127.6 (58.6%) 114.3 (55.1%) 79.5 (50.5%) 17.6 (7.6%) Other Liabilities 6.1 (2.2%) 7.7 (3.5%) 7.9 (3.8%) 5.6 (3.5%) 16.1 (7%)

Amounts in Billions (rounded). Source: http://www.ffec.gov/nicpubweb/nicweb


value of the bank holding company’s assets fell by an amount sufficient to extinguish this debt. This is unlikely. At a minimum, ongoing regulatory supervision of bank holding companies are likely to make such significant declines in assets values infrequent in the range of cases. In fact, the FDIC is also unlikely to be the residual claimant when the very largest banks fail. These mega-banks are the subsidiaries of the systematically important bank holding companies. As a result, the residual claimant principle suggests that the FDIC should have less control over the failure of large banks and their parent corporations, not more.241

Supporters of the OLA might argue that the FDIC’s relative lack of a financial interest makes it an impartial agent, much like a bankruptcy trustee in Chapter 7 or a state receiver.242 A crucial difference, however, is that bankruptcy and state receivership law subjects the trustee or receiver’s control to judicial oversight and grants interested parties the right to be heard. The FDIC is not subject to similar oversight.243 Judicial scrutiny helps mitigate the risk that the agent will slacken her efforts or serve her own interest. On the other hand, this oversight delays resolution and may increase administrative costs. It is at least theoretically possible that the benefits provided by speed would outweigh the costs of a loss of judicial oversight.

In Section III we rejected speed and secrecy as a justification for the FDIC’s control of bank resolutions, and we are skeptical about its use to justify FDIC control of very large bank holding companies or other financial concerns. The first reason we rejected speed and secrecy in the bank context does not apply to bank holding companies. Bank resolutions promptly reimburse insured depositors; they do not promptly dispose of the failed bank’s assets. Bank holding companies and nonbank financial companies do not take in deposits. However, they do have short-term obligations that may make them vulnerable if their counter-parties get nervous and try to quickly

241. This is subject to the same caveats that we expressed above. For example, the assignment of control to the FDIC might still be efficient even if it is not the residual claimant if this gives subordinated claims an incentive to ensure that the firm does not become insolvent and the FDIC does not seize control. 242. This assumes that the FDIC has no stake in the failed firm or that there is sufficient debt junior to the FDIC to insulate the FDIC from loss. 243. See supra note 100 and accompanying text (noting that claimants against failed banks do not, under current law, have recourse to voting procedures or judicial interference in disposition of the bank’s assets); cf. Dodd-Frank § 202(a)(1)(A)(iii), H.R. 4173 § 202(a)(1)(A)(iii) (stating that review of the Treasury Secretary’s determination of company actual or immanent default is subject to the "arbitrary and capricious" standard); id. § 210(c), H.R. 4173 § 210(c) (stating that court action which limits the FDIC’s exercise of power if permitted to the extent that it is provided in the OLA).

1050 67 WASH. & LEE L. REV. 985 (2010)

withdraw their funds. A prompt resolution could perhaps ease counter-party fears if it could somehow solve the liquidity crisis by quickly disposing of the assets and reimbursing creditors.

The FDIC’s experience in quickly resolving failed banks is not encouraging. It gives no reason to believe that the FDIC would resolve a bank holding company much more quickly than would bankruptcy, unless the FDIC were willing to provide substantial assistance that shifts much of the risk of loss to the FDIC itself.244 Recall that in most bank failures the FDIC retains at least 75% of the failed bank’s assets after the initial assignment and assumption245 and liquidates them over a period of about four years.246 By contrast, bankruptcy courts required a little less than one and a half years, on average, to dispose of the filings of large publicly traded corporations between 1995 and 2008.247 It is reasonable to expect an even slower pace in disposing of assets at the holding company level that affiliate banks do not typically maintain. Granted, the FDIC sometimes sells all or substantially all of the assets of the failed bank immediately upon seizing the failed bank. These resolutions are marginally quicker than the fastest bankruptcies resolved by use of Section 363. However, in conducting these immediate sales the FDIC usually agrees to bear much of the risk of a fall in the value of these assets through a loss-sharing agreement.248 It is the FDIC’s willingness

244. Sheila Bair, the chairman of the FDIC, might disagree with this assessment. In a recent editorial she argued in favor of the proposed reforms because the FDIC has a proven record of resolving failed banks without disrupting credit services. See Sheila Bair, Editorial, Beyond Bankruptcy and Bailouts, WALL ST. J., Apr. 5, 2010, at A19 (arguing that the FDIC’s record supports the proposed reforms). We argue that this ability to maintain liquidity for depositors depends on the FDIC’s deposit insurance and not its expertise in resolving failed financial institutions. Others are more skeptical of the value of the FDIC’s experience, noting that the banks resolved by the FDIC are much smaller and have qualitatively different assets and liabilities than the systematically important institutions that are the subject of the bill. See, e.g., Peter J. Wallison & David Skeel, Editorial, The Dodd Bill: Bailouts Forever, WALL ST. J., Apr. 7, 2010, at A15 (arguing against the proposed bill based on the nature of the systematically important institutions it targets). 245. See supra note 81 and accompanying text (noting that in 51% of reported cases, less than 25% of the failed bank’s assets were purchased back). 246. See supra note 13 and accompanying text (providing an average timeline for liquidation of the failed bank’s assets). 247. We used Lynn LoPucki’s Bankruptcy Research Database to calculate this figure. See Lynn M. LoPucki’s Bankruptcy Research Database, http://lopucki.law.ucla.edu/ bankruptcy_research.asp (last visited Sept. 28, 2010) (providing a database of bankruptcy cases for research) (on file with the Washington and Lee Law Review). 248. See supra notes 81–82 and accompanying text (noting that when the FDIC was able to sell the majority of the assets of the failed bank, it usually entered into a loss-sharing agreement with the buyer, retaining much of the risk of a decline in value).


to retain troubled assets or assume the risk of a decline in their value that provides liquidity and prevents the assets from being sold at "fire sale" prices. In a bank failure the FDIC’s willingness to assume this risk is unobjectionable; the FDIC already accepts this risk when it insures the major liability of the bank—the domestic deposits. The FDIC has not, however, guaranteed the short-term obligations of bank holding companies. Although bank failures cannot be directly compared to bankruptcy filings, the burden of proof is on the proponents of change in the resolution of bank holding companies.

VI. Conclusion

It is not obvious why the FDIC should have the central role it has in resolving failed banks. The bankruptcy law of most other countries does not give bank regulators this role. Two arguments sometimes are given to justify the FDIC’s control over the bank resolution process: speed in asset disposal and the FDIC’s status as the largest creditor of the failed bank. The argument from speed wrongly conflates the quickness with which assets are disposed with the quickness with which deposits can convert their deposits into cash. Speed in disposing of assets has nothing to do with preserving the liquidity of deposits. This Article, however, takes seriously the FDIC’s status as the largest creditor of the typical failed bank and the implications of this status. Data suggest that the likely asset values of most failed banks makes the FDIC the residual claimant on those assets. It therefore has the proper incentives to act to maximize these asset values. Other stakeholders in those assets do not have the same incentives. This is distinctive of most bank insolvencies, and justifies giving the FDIC control of the resolution process. Things could be different and sometimes are. The capital structures of mega-banks often differ from those of the typical failed bank. Bank holding companies too exhibit more complicated capital structures, with significant general and subordinated debt. With large banks, the allocation of control to the FDIC is presumptively unjustified. Nonregulatory claimants likely are the residual claimants on the assets of bank holding companies. In both cases the residual claimant principle demands that these claimants be given a voice. The FDIC’s control over the resolution of these entities must be justified on other grounds.

why banks are not allowed in bankruptcy - school of law

Documents